[core] Don't fail to queue on 409 responses

[VaguelySerious]

Noticed error logs from 409s falling through the queue, which are confusing for users. Log example:

Queue callback error: Error [WorkflowAPIError]: Cannot update workflow run wrun_01KKVYQ2TNA4PCEH9MA0SHN7DV with status 'completed'. Operation requires status 'running' or 'pending'.
at cP (.next/server/chunks/[root-of-the-server]__fd848f42._.js:63:41)
at async gA (.next/server/chunks/[root-of-the-server]__fd848f42._.js:67:3954)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:86:8)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:60:148157)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:60:146659)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:83:2206)
at async default (.next/server/chunks/[root-of-the-server]__fd848f42._.js:67:7813)
at async u7.processMessage (.next/server/chunks/[root-of-the-server]__fd848f42._.js:62:13992)
at async u7.consume (.next/server/chunks/[root-of-the-server]__fd848f42._.js:62:15197) {
cause: undefined,
status: 409,
code: undefined,
url: 'https://vercel-workflow.com/api/v1/runs/wrun_01KKVYQ2TNA4PCEH9MA0SHN7DV'
}

so I ensures we catch and gracefully return for these cases. I had Claude run an eval on whether this blocks any real use cases. Eval below:

---

Concurrent scenarios and why skipping is safe

Scenario 1: Two handlers race on run_started

Two queue messages for the same run arrive while it's still pending. Both handlers call world.runs.get(), see pending, and try to create a run_started event.

Handler A: runs.get() → pending → events.create(run_started) → ✅ succeeds → continues replay loop
Handler B: runs.get() → pending → events.create(run_started) → ❌ 409 → returns early

Why it's safe: Handler A transitions the run to running and enters the replay loop. The replay loop either completes the workflow or suspends it with queued continuations. Handler B's early exit is a no-op — Handler A guarantees progress.

Location:runtime.ts catch block after run_started event creation.

Scenario 2: Two handlers race on run_completed

Two concurrent replay loops both reach the end of the workflow and try to create run_completed.

Handler A: events.create(run_completed) → ✅ succeeds → returns
Handler B: events.create(run_completed) → ❌ 409 → logs info → returns

Why it's safe: The run is completed. Both handlers return. No continuation is needed.

Location:runtime.ts catch block after run_completed event creation.

Scenario 3: Two handlers race on run_failed

Same as Scenario 2 but for failure. Both handlers detect an error in user code and try to fail the run.

Handler A: events.create(run_failed) → ✅ succeeds → returns
Handler B: events.create(run_failed) → ❌ 409 → logs info → returns

Why it's safe: The run is failed. Both handlers return. No continuation is needed.

Location:runtime.ts catch block after run_failed event creation.

Scenario 4: Run completes while another handler creates hooks

Handler A completes the workflow. Handler B (from an earlier queue message) is still in the suspension handler creating hook events.

Handler A: events.create(run_completed) → ✅ run is now terminal
Handler B: events.create(hook_created) → ❌ 409 (run terminal) → logs info, continues
→ returns from handleSuspension with pendingSteps
→ tries to execute inline step
→ events.create(step_started) → ❌ 410 (run gone) → returns { type: 'gone' }
→ caller returns

Why it's safe: Handler A already completed the workflow. Handler B's hook creation fails, and if any steps were queued, the step executor receives 410 on step_started and exits gracefully. The 410 on step_started is the safety net — even if the suspension handler returns pending steps after a 409, the step executor won't make progress on a finished run.

Location:suspension-handler.ts catch blocks for hook_created and hook_disposed; step-executor.ts 410 handling on step_started.

Scenario 5: Two handlers race on wait_completed

During the replay loop, both handlers see an elapsed wait and try to complete it.

Handler A: events.create(wait_completed) → ✅ succeeds → adds event to cache → continues replay
Handler B: events.create(wait_completed) → ❌ 409 → continue (skip this wait) → continues replay

Why it's safe: Both handlers continue their replay loops. The event log is the same for both (the winning handler's wait_completed is visible to future reads). Subsequent replay from either handler converges to the same state because replay is deterministic and event-sourced.

Location:runtime.ts wait completion loop with continue on 409.

Scenario 6: Two handlers race on step_completed

Two handlers execute the same step concurrently (e.g., both received the step's queue message). Both finish execution and try to create step_completed.

Handler A: events.create(step_completed) → ✅ succeeds → queues workflow continuation → returns
Handler B: events.create(step_completed) → ❌ 409 → returns WITHOUT queuing continuation

Why it's safe: Handler A queues the workflow continuation. Handler B does not — this is correct because only one continuation should be queued per step completion. If both queued, there would be redundant replay (which is safe but wasteful).

Location:step-executor.ts catch on step_completed event creation.

Scenario 7: Two handlers race on step_started

Two handlers try to start the same step. Handler A wins; Handler B gets 409 because the step transitioned to a terminal state.

Handler A: events.create(step_started) → ✅ succeeds → executes step → queues continuation
Handler B: events.create(step_started) → ❌ 409 (step terminal) → returns { type: 'skipped' }
→ caller queues workflow continuation (runtime.ts replay loop)

Why it's safe: Both handlers ensure workflow continuation — Handler A via step completion, Handler B via the replay loop re-queuing the workflow. The redundant replay is safe because event-sourced replay is convergent.

Location:step-executor.ts 409 handling on step_started.

Scenario 8: step_created/wait_created duplicate

During suspension handling, two concurrent replays both try to create the same step or wait event.

Handler A: events.create(step_created) → ✅ succeeds
Handler B: events.create(step_created) → ❌ 409 → logs info → continues

Why it's safe: The step/wait entity already exists. The suspension handler continues processing other items. The step will be executed by whichever handler picks it up from the queue.

Location:suspension-handler.ts catch blocks for step_created and wait_created.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
example-nextjs-workflow-turbopack	Ready	Preview, Comment	Mar 17, 2026 5:00pm
example-nextjs-workflow-webpack	Ready	Preview, Comment	Mar 17, 2026 5:00pm
example-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-astro-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-express-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-fastify-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-hono-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-nitro-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-nuxt-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-sveltekit-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-vite-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workflow-docs	Ready	Preview, Comment, Open in v0	Mar 17, 2026 5:00pm
workflow-nest	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workflow-swc-playground	Ready	Preview, Comment	Mar 17, 2026 5:00pm

[changeset-bot]

🦋 Changeset detected

Latest commit: 9668d09

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 16 packages

Name	Type
@workflow/core	Patch
@workflow/builders	Patch
@workflow/cli	Patch
@workflow/next	Patch
@workflow/nitro	Patch
@workflow/vitest	Patch
@workflow/web-shared	Patch
workflow	Patch
@workflow/world-testing	Patch
@workflow/astro	Patch
@workflow/nest	Patch
@workflow/rollup	Patch
@workflow/sveltekit	Patch
@workflow/vite	Patch
@workflow/nuxt	Patch
@workflow/ai	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

🧪 E2E Test Results

❌ Some tests failed

Summary

	Passed	Failed	Skipped	Total
✅ ▲ Vercel Production	747	0	67	814
✅ 💻 Local Development	770	0	118	888
✅ 📦 Local Production	770	0	118	888
✅ 🐘 Local Postgres	770	0	118	888
✅ 🪟 Windows	71	0	3	74
❌ 🌍 Community Worlds	116	55	15	186
✅ 📋 Other	195	0	27	222
Total	3439	55	466	3960

❌ Failed Tests

🌍 Community Worlds (55 failed)

mongodb (3 failed):

• hookWorkflow is not resumable via public webhook endpoint
• webhookWorkflow
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously

redis (2 failed):

• hookWorkflow is not resumable via public webhook endpoint
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously

turso (50 failed):

• addTenWorkflow
• addTenWorkflow
• wellKnownAgentWorkflow (.well-known/agent)
• should work with react rendering in step
• promiseAllWorkflow
• promiseRaceWorkflow
• promiseAnyWorkflow
• importedStepOnlyWorkflow
• hookWorkflow
• hookWorkflow is not resumable via public webhook endpoint
• webhookWorkflow
• sleepingWorkflow
• parallelSleepWorkflow
• nullByteWorkflow
• workflowAndStepMetadataWorkflow
• fetchWorkflow
• promiseRaceStressTestWorkflow
• error handling error propagation workflow errors nested function calls preserve message and stack trace
• error handling error propagation workflow errors cross-file imports preserve message and stack trace
• error handling error propagation step errors basic step error preserves message and stack trace
• error handling error propagation step errors cross-file step error preserves message and function names in stack
• error handling retry behavior regular Error retries until success
• error handling retry behavior FatalError fails immediately without retries
• error handling retry behavior RetryableError respects custom retryAfter delay
• error handling retry behavior maxRetries=0 disables retries
• error handling catchability FatalError can be caught and detected with FatalError.is()
• hookCleanupTestWorkflow - hook token reuse after workflow completion
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously
• hookDisposeTestWorkflow - hook token reuse after explicit disposal while workflow still running
• stepFunctionPassingWorkflow - step function references can be passed as arguments (without closure vars)
• stepFunctionWithClosureWorkflow - step function with closure variables passed as argument
• closureVariableWorkflow - nested step functions with closure variables
• spawnWorkflowFromStepWorkflow - spawning a child workflow using start() inside a step
• health check (queue-based) - workflow and step endpoints respond to health check messages
• pathsAliasWorkflow - TypeScript path aliases resolve correctly
• Calculator.calculate - static workflow method using static step methods from another class
• AllInOneService.processNumber - static workflow method using sibling static step methods
• ChainableService.processWithThis - static step methods using this to reference the class
• thisSerializationWorkflow - step function invoked with .call() and .apply()
• customSerializationWorkflow - custom class serialization with WORKFLOW_SERIALIZE/WORKFLOW_DESERIALIZE
• instanceMethodStepWorkflow - instance methods with "use step" directive
• crossContextSerdeWorkflow - classes defined in step code are deserializable in workflow context
• stepFunctionAsStartArgWorkflow - step function reference passed as start() argument
• cancelRun - cancelling a running workflow
• cancelRun via CLI - cancelling a running workflow
• pages router addTenWorkflow via pages router
• pages router promiseAllWorkflow via pages router
• pages router sleepingWorkflow via pages router
• hookWithSleepWorkflow - hook payloads delivered correctly with concurrent sleep
• sleepWithSequentialStepsWorkflow - sequential steps work with concurrent sleep (control)

Details by Category

✅ ▲ Vercel Production

App	Passed	Failed	Skipped
✅ astro	67	0	7
✅ example	67	0	7
✅ express	67	0	7
✅ fastify	67	0	7
✅ hono	67	0	7
✅ nextjs-turbopack	72	0	2
✅ nextjs-webpack	72	0	2
✅ nitro	67	0	7
✅ nuxt	67	0	7
✅ sveltekit	67	0	7
✅ vite	67	0	7

✅ 💻 Local Development

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 📦 Local Production

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 🐘 Local Postgres

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 🪟 Windows

App	Passed	Failed	Skipped
✅ nextjs-turbopack	71	0	3

❌ 🌍 Community Worlds

App	Passed	Failed	Skipped
✅ mongodb-dev	3	0	2
❌ mongodb	51	3	3
✅ redis-dev	3	0	2
❌ redis	52	2	3
✅ turso-dev	3	0	2
❌ turso	4	50	3

✅ 📋 Other

App	Passed	Failed	Skipped
✅ e2e-local-dev-nest-stable	65	0	9
✅ e2e-local-postgres-nest-stable	65	0	9
✅ e2e-local-prod-nest-stable	65	0	9

---

📋 View full workflow run

[github-actions]

📊 Benchmark Results

📈 Comparing against baseline from main branch. Green 🟢 = faster, Red 🔺 = slower.

workflow with no steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	0.038s (-15.7% 🟢)	1.005s (~)	0.968s	10	1.00x
💻 Local	Nitro	0.046s (+6.7% 🔺)	1.006s (~)	0.959s	10	1.23x
💻 Local	Next.js (Turbopack)	0.051s	1.006s	0.955s	10	1.36x
🌐 Redis	Next.js (Turbopack)	0.054s	1.005s	0.951s	10	1.44x
🐘 Postgres	Nitro	0.059s (-4.7%)	1.011s (~)	0.952s	10	1.56x
🐘 Postgres	Express	0.059s (-14.8% 🟢)	1.011s (~)	0.952s	10	1.57x
🌐 MongoDB	Next.js (Turbopack)	0.079s	1.008s	0.929s	10	2.11x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	0.465s (-33.5% 🟢)	2.036s (-22.7% 🟢)	1.571s	10	1.00x
▲ Vercel	Express	0.504s (-20.0% 🟢)	2.286s (-14.4% 🟢)	1.781s	10	1.08x
▲ Vercel	Next.js (Turbopack)	0.511s (-17.6% 🟢)	2.559s (+13.2% 🔺)	2.049s	10	1.10x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

workflow with 1 step

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.093s (-3.3%)	2.006s (~)	0.912s	10	1.00x
🌐 Redis	Next.js (Turbopack)	1.122s	2.006s	0.885s	10	1.03x
💻 Local	Nitro	1.130s (~)	2.006s (~)	0.877s	10	1.03x
💻 Local	Next.js (Turbopack)	1.130s	2.006s	0.876s	10	1.03x
🐘 Postgres	Express	1.153s (+0.7%)	2.014s (~)	0.860s	10	1.05x
🐘 Postgres	Nitro	1.158s (+1.1%)	2.012s (~)	0.855s	10	1.06x
🌐 MongoDB	Next.js (Turbopack)	1.307s	2.008s	0.700s	10	1.20x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.040s (-3.0%)	3.224s (-4.4%)	1.184s	10	1.00x
▲ Vercel	Express	2.053s (+0.9%)	3.735s (+12.4% 🔺)	1.682s	10	1.01x
▲ Vercel	Next.js (Turbopack)	2.168s (-2.0%)	3.924s (+13.8% 🔺)	1.756s	10	1.06x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

workflow with 10 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	10.615s (-2.8%)	11.023s (~)	0.408s	3	1.00x
🌐 Redis	Next.js (Turbopack)	10.784s	11.023s	0.239s	3	1.02x
💻 Local	Next.js (Turbopack)	10.812s	11.023s	0.211s	3	1.02x
🐘 Postgres	Nitro	10.927s (~)	11.042s (~)	0.115s	3	1.03x
🐘 Postgres	Express	10.970s (~)	11.375s (~)	0.405s	3	1.03x
💻 Local	Nitro	10.976s (+0.8%)	11.024s (~)	0.048s	3	1.03x
🌐 MongoDB	Next.js (Turbopack)	12.203s	13.015s	0.812s	3	1.15x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	16.779s (-7.4% 🟢)	18.655s (-5.6% 🟢)	1.876s	2	1.00x
▲ Vercel	Nitro	16.868s (+1.3%)	18.197s (-1.3%)	1.329s	2	1.01x
▲ Vercel	Next.js (Turbopack)	17.826s (+0.9%)	19.468s (+1.2%)	1.641s	2	1.06x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

workflow with 25 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	26.743s (-3.1%)	27.051s (-3.6%)	0.308s	3	1.00x
🌐 Redis	Next.js (Turbopack)	26.760s	27.049s	0.289s	3	1.00x
💻 Local	Next.js (Turbopack)	27.128s	28.053s	0.925s	3	1.01x
🐘 Postgres	Express	27.240s (~)	28.066s (~)	0.826s	3	1.02x
🐘 Postgres	Nitro	27.277s (~)	28.065s (~)	0.788s	3	1.02x
💻 Local	Nitro	27.583s (~)	28.053s (~)	0.470s	3	1.03x
🌐 MongoDB	Next.js (Turbopack)	30.482s	31.045s	0.564s	2	1.14x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	46.028s (+2.9%)	47.737s (+4.3%)	1.709s	2	1.00x
▲ Vercel	Next.js (Turbopack)	50.377s (+6.6% 🔺)	52.253s (+8.0% 🔺)	1.877s	2	1.09x
▲ Vercel	Express	53.606s (+21.0% 🔺)	55.319s (+22.1% 🔺)	1.713s	2	1.16x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

workflow with 50 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	53.541s	54.093s	0.552s	2	1.00x
🐘 Postgres	Express	54.300s (~)	55.094s (~)	0.793s	2	1.01x
🐘 Postgres	Nitro	54.476s (~)	55.105s (~)	0.629s	2	1.02x
💻 Local	Express	54.851s (-3.3%)	55.101s (-3.5%)	0.250s	2	1.02x
💻 Local	Next.js (Turbopack)	55.951s	56.100s	0.149s	2	1.05x
💻 Local	Nitro	56.657s (~)	57.106s (~)	0.450s	2	1.06x
🌐 MongoDB	Next.js (Turbopack)	60.675s	61.065s	0.390s	2	1.13x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	94.771s (-4.4%)	96.975s (-4.2%)	2.204s	1	1.00x
▲ Vercel	Nitro	96.605s (~)	98.384s (~)	1.779s	1	1.02x
▲ Vercel	Next.js (Turbopack)	98.753s (-3.9%)	100.796s (-2.8%)	2.043s	1	1.04x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

Promise.all with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	1.345s	2.006s	0.662s	15	1.00x
🐘 Postgres	Express	1.366s (-4.0%)	2.011s (~)	0.645s	15	1.02x
💻 Local	Express	1.467s (-5.2% 🟢)	2.006s (~)	0.538s	15	1.09x
💻 Local	Nitro	1.498s (-1.1%)	2.005s (~)	0.507s	15	1.11x
💻 Local	Next.js (Turbopack)	1.547s	2.006s	0.459s	15	1.15x
🌐 MongoDB	Next.js (Turbopack)	2.146s	3.009s	0.863s	10	1.60x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-
🐘 Postgres	Nitro	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.312s (-4.2%)	3.407s (-4.8%)	1.095s	9	1.00x
▲ Vercel	Next.js (Turbopack)	2.641s (-1.3%)	4.038s (+7.5% 🔺)	1.397s	8	1.14x
▲ Vercel	Express	2.648s (+5.0%)	4.138s (+13.0% 🔺)	1.490s	8	1.15x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Promise.all with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	2.567s	3.008s	0.440s	10	1.00x
🐘 Postgres	Nitro	2.595s (~)	3.014s (~)	0.419s	10	1.01x
💻 Local	Express	2.600s (-14.1% 🟢)	3.007s (-15.6% 🟢)	0.407s	10	1.01x
🐘 Postgres	Express	2.615s (~)	3.015s (~)	0.399s	10	1.02x
💻 Local	Next.js (Turbopack)	3.041s	3.760s	0.718s	8	1.18x
💻 Local	Nitro	3.121s (+7.5% 🔺)	3.565s (+11.1% 🔺)	0.444s	9	1.22x
🌐 MongoDB	Next.js (Turbopack)	4.747s	5.179s	0.432s	6	1.85x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Next.js (Turbopack)	2.799s (+7.7% 🔺)	4.348s (+21.1% 🔺)	1.550s	8	1.00x
▲ Vercel	Nitro	3.520s (+30.2% 🔺)	4.708s (+23.4% 🔺)	1.188s	7	1.26x
▲ Vercel	Express	3.663s (+47.9% 🔺)	5.183s (+40.8% 🔺)	1.520s	8	1.31x

🔍 Observability: Next.js (Turbopack) | Nitro | Express

Promise.all with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	4.027s (+1.4%)	4.588s (+3.2%)	0.561s	7	1.00x
🐘 Postgres	Nitro	4.032s (+0.7%)	4.590s (+3.1%)	0.558s	7	1.00x
🌐 Redis	Next.js (Turbopack)	4.081s	5.012s	0.931s	6	1.01x
💻 Local	Express	6.787s (-15.7% 🟢)	7.014s (-20.1% 🟢)	0.227s	5	1.69x
💻 Local	Next.js (Turbopack)	7.888s	8.517s	0.629s	4	1.96x
💻 Local	Nitro	8.207s (-1.6%)	8.521s (-5.5% 🟢)	0.314s	4	2.04x
🌐 MongoDB	Next.js (Turbopack)	10.067s	10.684s	0.617s	3	2.50x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Next.js (Turbopack)	5.520s (+44.3% 🔺)	7.146s (+37.1% 🔺)	1.626s	5	1.00x
▲ Vercel	Nitro	7.009s (+153.3% 🔺)	8.352s (+119.6% 🔺)	1.343s	4	1.27x
▲ Vercel	Express	10.371s (+264.0% 🔺)	12.346s (+209.2% 🔺)	1.975s	3	1.88x

🔍 Observability: Next.js (Turbopack) | Nitro | Express

Promise.race with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	1.303s	2.006s	0.703s	15	1.00x
🐘 Postgres	Nitro	1.422s (-1.3%)	2.011s (~)	0.589s	15	1.09x
🐘 Postgres	Express	1.442s (-1.5%)	2.011s (-3.2%)	0.570s	15	1.11x
💻 Local	Express	1.477s (-2.9%)	2.005s (~)	0.528s	15	1.13x
💻 Local	Nitro	1.523s (-1.1%)	2.006s (~)	0.483s	15	1.17x
💻 Local	Next.js (Turbopack)	1.564s	2.073s	0.509s	15	1.20x
🌐 MongoDB	Next.js (Turbopack)	2.174s	3.009s	0.835s	10	1.67x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.131s (-0.9%)	3.290s (-1.5%)	1.159s	10	1.00x
▲ Vercel	Next.js (Turbopack)	2.268s (-27.5% 🟢)	3.823s (-24.1% 🟢)	1.555s	8	1.06x
▲ Vercel	Express	2.591s (+17.2% 🔺)	4.029s (+17.2% 🔺)	1.438s	8	1.22x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Promise.race with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	2.526s (-3.2%)	3.012s (~)	0.486s	10	1.00x
🌐 Redis	Next.js (Turbopack)	2.532s	3.008s	0.477s	10	1.00x
🐘 Postgres	Nitro	2.730s (+7.4% 🔺)	3.015s (~)	0.285s	10	1.08x
💻 Local	Express	2.797s (-9.8% 🟢)	3.108s (-17.3% 🟢)	0.311s	10	1.11x
💻 Local	Next.js (Turbopack)	3.030s	3.760s	0.729s	8	1.20x
💻 Local	Nitro	3.124s (-1.8%)	3.886s (~)	0.762s	8	1.24x
🌐 MongoDB	Next.js (Turbopack)	4.728s	5.177s	0.449s	6	1.87x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.270s (-29.0% 🟢)	3.269s (-37.5% 🟢)	0.998s	10	1.00x
▲ Vercel	Express	2.378s (-32.1% 🟢)	3.827s (-17.2% 🟢)	1.449s	8	1.05x
▲ Vercel	Next.js (Turbopack)	3.093s (-5.5% 🟢)	4.610s (~)	1.517s	7	1.36x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

Promise.race with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	3.975s (-3.5%)	4.590s (~)	0.615s	7	1.00x
🐘 Postgres	Express	4.039s (+2.3%)	4.590s (+3.2%)	0.551s	7	1.02x
🌐 Redis	Next.js (Turbopack)	4.198s	4.725s	0.527s	7	1.06x
💻 Local	Express	7.733s (-14.1% 🟢)	8.267s (-10.8% 🟢)	0.533s	4	1.95x
💻 Local	Next.js (Turbopack)	8.255s	8.766s	0.512s	4	2.08x
💻 Local	Nitro	8.519s (+0.8%)	9.275s (+2.8%)	0.756s	4	2.14x
🌐 MongoDB	Next.js (Turbopack)	10.032s	10.680s	0.648s	3	2.52x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.724s (-27.3% 🟢)	3.647s (-28.7% 🟢)	0.923s	9	1.00x
▲ Vercel	Next.js (Turbopack)	3.571s (-3.3%)	5.369s (+9.7% 🔺)	1.798s	6	1.31x
▲ Vercel	Express	3.789s (+1.4%)	5.100s (-0.7%)	1.311s	6	1.39x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Stream Benchmarks(includes TTFB metrics)

workflow with stream

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	0.138s (-32.2% 🟢)	1.003s (~)	0.009s (-21.6% 🟢)	1.015s (~)	0.877s	10	1.00x
💻 Local	Next.js (Turbopack)	0.170s	1.002s	0.011s	1.017s	0.847s	10	1.24x
🌐 Redis	Next.js (Turbopack)	0.182s	1.000s	0.002s	1.007s	0.825s	10	1.33x
💻 Local	Nitro	0.212s (+7.3% 🔺)	1.003s (~)	0.011s (-2.7%)	1.017s (~)	0.804s	10	1.54x
🐘 Postgres	Express	0.215s (+2.8%)	0.996s (~)	0.002s (+15.4% 🔺)	1.013s (~)	0.797s	10	1.57x
🐘 Postgres	Nitro	0.246s (+9.4% 🔺)	0.993s (~)	0.002s (+41.7% 🔺)	1.014s (~)	0.767s	10	1.79x
🌐 MongoDB	Next.js (Turbopack)	0.474s	0.979s	0.002s	1.009s	0.534s	10	3.45x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	1.606s (+1.9%)	2.405s (-8.7% 🟢)	0.006s (-26.2% 🟢)	3.035s (-2.2%)	1.429s	10	1.00x
▲ Vercel	Nitro	1.658s (+1.2%)	2.506s (-14.3% 🟢)	0.504s (+8896.4% 🔺)	3.490s (+1.5%)	1.833s	10	1.03x
▲ Vercel	Next.js (Turbopack)	1.785s (-4.5%)	3.025s (+6.5% 🔺)	0.011s (+160.5% 🔺)	3.640s (+6.9% 🔺)	1.855s	10	1.11x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

Summary

Fastest Framework by World

Winner determined by most benchmark wins

World	🥇 Fastest Framework	Wins
💻 Local	Express	12/12
🐘 Postgres	Express	7/12
▲ Vercel	Nitro	7/12

Fastest World by Framework

Winner determined by most benchmark wins

Framework	🥇 Fastest World	Wins
Express	💻 Local	6/12
Next.js (Turbopack)	🌐 Redis	9/12
Nitro	🐘 Postgres	6/12

Column Definitions

• Workflow Time: Runtime reported by workflow (completedAt - createdAt) - primary metric
• TTFB: Time to First Byte - time from workflow start until first stream byte received (stream benchmarks only)
• Slurp: Time from first byte to complete stream consumption (stream benchmarks only)
• Wall Time: Total testbench time (trigger workflow + poll for result)
• Overhead: Testbench overhead (Wall Time - Workflow Time)
• Samples: Number of benchmark iterations run
• vs Fastest: How much slower compared to the fastest configuration for this benchmark

Worlds:

• 💻 Local: In-memory filesystem world (local development)
• 🐘 Postgres: PostgreSQL database world (local development)
• ▲ Vercel: Vercel production/preview deployment
• 🌐 Turso: Community world (local development)
• 🌐 MongoDB: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Jazz: Community world (local development)

---

📋 View full workflow run

[pranaygp]

Suggested change

	'Run already finished during setup, skipping',
	'Run already finished during setup, skipping replay',

maybe this would be clearer to the user? (assuming the codepath is only for run replays)

[pranaygp]

or "skipping redundant workflow execution"

[pranaygp]

human: LGTM! great detailed state check and ty for posting that on github PR review comment

[VaguelySerious]

Noticed error logs from 409s falling through the queue, which are confusing for users. Log example:

Queue callback error: Error [WorkflowAPIError]: Cannot update workflow run wrun_01KKVYQ2TNA4PCEH9MA0SHN7DV with status 'completed'. Operation requires status 'running' or 'pending'.
at cP (.next/server/chunks/[root-of-the-server]__fd848f42._.js:63:41)
at async gA (.next/server/chunks/[root-of-the-server]__fd848f42._.js:67:3954)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:86:8)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:60:148157)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:60:146659)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:83:2206)
at async default (.next/server/chunks/[root-of-the-server]__fd848f42._.js:67:7813)
at async u7.processMessage (.next/server/chunks/[root-of-the-server]__fd848f42._.js:62:13992)
at async u7.consume (.next/server/chunks/[root-of-the-server]__fd848f42._.js:62:15197) {
cause: undefined,
status: 409,
code: undefined,
url: 'https://vercel-workflow.com/api/v1/runs/wrun_01KKVYQ2TNA4PCEH9MA0SHN7DV'
}

so I ensures we catch and gracefully return for these cases. I had Claude run an eval on whether this blocks any real use cases. Eval below:

---

Concurrent scenarios and why skipping is safe

Scenario 1: Two handlers race on run_started

Two queue messages for the same run arrive while it's still pending. Both handlers call world.runs.get(), see pending, and try to create a run_started event.

Handler A: runs.get() → pending → events.create(run_started) → ✅ succeeds → continues replay loop
Handler B: runs.get() → pending → events.create(run_started) → ❌ 409 → returns early

Why it's safe: Handler A transitions the run to running and enters the replay loop. The replay loop either completes the workflow or suspends it with queued continuations. Handler B's early exit is a no-op — Handler A guarantees progress.

Location:runtime.ts catch block after run_started event creation.

Scenario 2: Two handlers race on run_completed

Two concurrent replay loops both reach the end of the workflow and try to create run_completed.

Handler A: events.create(run_completed) → ✅ succeeds → returns
Handler B: events.create(run_completed) → ❌ 409 → logs info → returns

Why it's safe: The run is completed. Both handlers return. No continuation is needed.

Location:runtime.ts catch block after run_completed event creation.

Scenario 3: Two handlers race on run_failed

Same as Scenario 2 but for failure. Both handlers detect an error in user code and try to fail the run.

Handler A: events.create(run_failed) → ✅ succeeds → returns
Handler B: events.create(run_failed) → ❌ 409 → logs info → returns

Why it's safe: The run is failed. Both handlers return. No continuation is needed.

Location:runtime.ts catch block after run_failed event creation.

Scenario 4: Run completes while another handler creates hooks

Handler A completes the workflow. Handler B (from an earlier queue message) is still in the suspension handler creating hook events.

Handler A: events.create(run_completed) → ✅ run is now terminal
Handler B: events.create(hook_created) → ❌ 409 (run terminal) → logs info, continues
→ returns from handleSuspension with pendingSteps
→ tries to execute inline step
→ events.create(step_started) → ❌ 410 (run gone) → returns { type: 'gone' }
→ caller returns

Why it's safe: Handler A already completed the workflow. Handler B's hook creation fails, and if any steps were queued, the step executor receives 410 on step_started and exits gracefully. The 410 on step_started is the safety net — even if the suspension handler returns pending steps after a 409, the step executor won't make progress on a finished run.

Location:suspension-handler.ts catch blocks for hook_created and hook_disposed; step-executor.ts 410 handling on step_started.

Scenario 5: Two handlers race on wait_completed

During the replay loop, both handlers see an elapsed wait and try to complete it.

Handler A: events.create(wait_completed) → ✅ succeeds → adds event to cache → continues replay
Handler B: events.create(wait_completed) → ❌ 409 → continue (skip this wait) → continues replay

Why it's safe: Both handlers continue their replay loops. The event log is the same for both (the winning handler's wait_completed is visible to future reads). Subsequent replay from either handler converges to the same state because replay is deterministic and event-sourced.

Location:runtime.ts wait completion loop with continue on 409.

Scenario 6: Two handlers race on step_completed

Two handlers execute the same step concurrently (e.g., both received the step's queue message). Both finish execution and try to create step_completed.

Handler A: events.create(step_completed) → ✅ succeeds → queues workflow continuation → returns
Handler B: events.create(step_completed) → ❌ 409 → returns WITHOUT queuing continuation

Why it's safe: Handler A queues the workflow continuation. Handler B does not — this is correct because only one continuation should be queued per step completion. If both queued, there would be redundant replay (which is safe but wasteful).

Location:step-executor.ts catch on step_completed event creation.

Scenario 7: Two handlers race on step_started

Two handlers try to start the same step. Handler A wins; Handler B gets 409 because the step transitioned to a terminal state.

Handler A: events.create(step_started) → ✅ succeeds → executes step → queues continuation
Handler B: events.create(step_started) → ❌ 409 (step terminal) → returns { type: 'skipped' }
→ caller queues workflow continuation (runtime.ts replay loop)

Why it's safe: Both handlers ensure workflow continuation — Handler A via step completion, Handler B via the replay loop re-queuing the workflow. The redundant replay is safe because event-sourced replay is convergent.

Location:step-executor.ts 409 handling on step_started.

Scenario 8: step_created/wait_created duplicate

During suspension handling, two concurrent replays both try to create the same step or wait event.

Handler A: events.create(step_created) → ✅ succeeds
Handler B: events.create(step_created) → ❌ 409 → logs info → continues

Why it's safe: The step/wait entity already exists. The suspension handler continues processing other items. The step will be executed by whichever handler picks it up from the queue.

Location:suspension-handler.ts catch blocks for step_created and wait_created.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
example-nextjs-workflow-turbopack	Ready	Preview, Comment	Mar 17, 2026 5:00pm
example-nextjs-workflow-webpack	Ready	Preview, Comment	Mar 17, 2026 5:00pm
example-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-astro-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-express-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-fastify-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-hono-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-nitro-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-nuxt-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-sveltekit-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-vite-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workflow-docs	Ready	Preview, Comment, Open in v0	Mar 17, 2026 5:00pm
workflow-nest	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workflow-swc-playground	Ready	Preview, Comment	Mar 17, 2026 5:00pm

[changeset-bot]

🦋 Changeset detected

Latest commit: 9668d09

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 16 packages

Name	Type
@workflow/core	Patch
@workflow/builders	Patch
@workflow/cli	Patch
@workflow/next	Patch
@workflow/nitro	Patch
@workflow/vitest	Patch
@workflow/web-shared	Patch
workflow	Patch
@workflow/world-testing	Patch
@workflow/astro	Patch
@workflow/nest	Patch
@workflow/rollup	Patch
@workflow/sveltekit	Patch
@workflow/vite	Patch
@workflow/nuxt	Patch
@workflow/ai	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

🧪 E2E Test Results

❌ Some tests failed

Summary

	Passed	Failed	Skipped	Total
✅ ▲ Vercel Production	747	0	67	814
✅ 💻 Local Development	770	0	118	888
✅ 📦 Local Production	770	0	118	888
✅ 🐘 Local Postgres	770	0	118	888
✅ 🪟 Windows	71	0	3	74
❌ 🌍 Community Worlds	116	55	15	186
✅ 📋 Other	195	0	27	222
Total	3439	55	466	3960

❌ Failed Tests

🌍 Community Worlds (55 failed)

mongodb (3 failed):

• hookWorkflow is not resumable via public webhook endpoint
• webhookWorkflow
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously

redis (2 failed):

• hookWorkflow is not resumable via public webhook endpoint
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously

turso (50 failed):

• addTenWorkflow
• addTenWorkflow
• wellKnownAgentWorkflow (.well-known/agent)
• should work with react rendering in step
• promiseAllWorkflow
• promiseRaceWorkflow
• promiseAnyWorkflow
• importedStepOnlyWorkflow
• hookWorkflow
• hookWorkflow is not resumable via public webhook endpoint
• webhookWorkflow
• sleepingWorkflow
• parallelSleepWorkflow
• nullByteWorkflow
• workflowAndStepMetadataWorkflow
• fetchWorkflow
• promiseRaceStressTestWorkflow
• error handling error propagation workflow errors nested function calls preserve message and stack trace
• error handling error propagation workflow errors cross-file imports preserve message and stack trace
• error handling error propagation step errors basic step error preserves message and stack trace
• error handling error propagation step errors cross-file step error preserves message and function names in stack
• error handling retry behavior regular Error retries until success
• error handling retry behavior FatalError fails immediately without retries
• error handling retry behavior RetryableError respects custom retryAfter delay
• error handling retry behavior maxRetries=0 disables retries
• error handling catchability FatalError can be caught and detected with FatalError.is()
• hookCleanupTestWorkflow - hook token reuse after workflow completion
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously
• hookDisposeTestWorkflow - hook token reuse after explicit disposal while workflow still running
• stepFunctionPassingWorkflow - step function references can be passed as arguments (without closure vars)
• stepFunctionWithClosureWorkflow - step function with closure variables passed as argument
• closureVariableWorkflow - nested step functions with closure variables
• spawnWorkflowFromStepWorkflow - spawning a child workflow using start() inside a step
• health check (queue-based) - workflow and step endpoints respond to health check messages
• pathsAliasWorkflow - TypeScript path aliases resolve correctly
• Calculator.calculate - static workflow method using static step methods from another class
• AllInOneService.processNumber - static workflow method using sibling static step methods
• ChainableService.processWithThis - static step methods using this to reference the class
• thisSerializationWorkflow - step function invoked with .call() and .apply()
• customSerializationWorkflow - custom class serialization with WORKFLOW_SERIALIZE/WORKFLOW_DESERIALIZE
• instanceMethodStepWorkflow - instance methods with "use step" directive
• crossContextSerdeWorkflow - classes defined in step code are deserializable in workflow context
• stepFunctionAsStartArgWorkflow - step function reference passed as start() argument
• cancelRun - cancelling a running workflow
• cancelRun via CLI - cancelling a running workflow
• pages router addTenWorkflow via pages router
• pages router promiseAllWorkflow via pages router
• pages router sleepingWorkflow via pages router
• hookWithSleepWorkflow - hook payloads delivered correctly with concurrent sleep
• sleepWithSequentialStepsWorkflow - sequential steps work with concurrent sleep (control)

Details by Category

✅ ▲ Vercel Production

App	Passed	Failed	Skipped
✅ astro	67	0	7
✅ example	67	0	7
✅ express	67	0	7
✅ fastify	67	0	7
✅ hono	67	0	7
✅ nextjs-turbopack	72	0	2
✅ nextjs-webpack	72	0	2
✅ nitro	67	0	7
✅ nuxt	67	0	7
✅ sveltekit	67	0	7
✅ vite	67	0	7

✅ 💻 Local Development

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 📦 Local Production

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 🐘 Local Postgres

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 🪟 Windows

App	Passed	Failed	Skipped
✅ nextjs-turbopack	71	0	3

❌ 🌍 Community Worlds

App	Passed	Failed	Skipped
✅ mongodb-dev	3	0	2
❌ mongodb	51	3	3
✅ redis-dev	3	0	2
❌ redis	52	2	3
✅ turso-dev	3	0	2
❌ turso	4	50	3

✅ 📋 Other

App	Passed	Failed	Skipped
✅ e2e-local-dev-nest-stable	65	0	9
✅ e2e-local-postgres-nest-stable	65	0	9
✅ e2e-local-prod-nest-stable	65	0	9

---

📋 View full workflow run

[github-actions]

📊 Benchmark Results

📈 Comparing against baseline from main branch. Green 🟢 = faster, Red 🔺 = slower.

workflow with no steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	0.038s (-15.7% 🟢)	1.005s (~)	0.968s	10	1.00x
💻 Local	Nitro	0.046s (+6.7% 🔺)	1.006s (~)	0.959s	10	1.23x
💻 Local	Next.js (Turbopack)	0.051s	1.006s	0.955s	10	1.36x
🌐 Redis	Next.js (Turbopack)	0.054s	1.005s	0.951s	10	1.44x
🐘 Postgres	Nitro	0.059s (-4.7%)	1.011s (~)	0.952s	10	1.56x
🐘 Postgres	Express	0.059s (-14.8% 🟢)	1.011s (~)	0.952s	10	1.57x
🌐 MongoDB	Next.js (Turbopack)	0.079s	1.008s	0.929s	10	2.11x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	0.465s (-33.5% 🟢)	2.036s (-22.7% 🟢)	1.571s	10	1.00x
▲ Vercel	Express	0.504s (-20.0% 🟢)	2.286s (-14.4% 🟢)	1.781s	10	1.08x
▲ Vercel	Next.js (Turbopack)	0.511s (-17.6% 🟢)	2.559s (+13.2% 🔺)	2.049s	10	1.10x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

workflow with 1 step

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.093s (-3.3%)	2.006s (~)	0.912s	10	1.00x
🌐 Redis	Next.js (Turbopack)	1.122s	2.006s	0.885s	10	1.03x
💻 Local	Nitro	1.130s (~)	2.006s (~)	0.877s	10	1.03x
💻 Local	Next.js (Turbopack)	1.130s	2.006s	0.876s	10	1.03x
🐘 Postgres	Express	1.153s (+0.7%)	2.014s (~)	0.860s	10	1.05x
🐘 Postgres	Nitro	1.158s (+1.1%)	2.012s (~)	0.855s	10	1.06x
🌐 MongoDB	Next.js (Turbopack)	1.307s	2.008s	0.700s	10	1.20x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.040s (-3.0%)	3.224s (-4.4%)	1.184s	10	1.00x
▲ Vercel	Express	2.053s (+0.9%)	3.735s (+12.4% 🔺)	1.682s	10	1.01x
▲ Vercel	Next.js (Turbopack)	2.168s (-2.0%)	3.924s (+13.8% 🔺)	1.756s	10	1.06x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

workflow with 10 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	10.615s (-2.8%)	11.023s (~)	0.408s	3	1.00x
🌐 Redis	Next.js (Turbopack)	10.784s	11.023s	0.239s	3	1.02x
💻 Local	Next.js (Turbopack)	10.812s	11.023s	0.211s	3	1.02x
🐘 Postgres	Nitro	10.927s (~)	11.042s (~)	0.115s	3	1.03x
🐘 Postgres	Express	10.970s (~)	11.375s (~)	0.405s	3	1.03x
💻 Local	Nitro	10.976s (+0.8%)	11.024s (~)	0.048s	3	1.03x
🌐 MongoDB	Next.js (Turbopack)	12.203s	13.015s	0.812s	3	1.15x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	16.779s (-7.4% 🟢)	18.655s (-5.6% 🟢)	1.876s	2	1.00x
▲ Vercel	Nitro	16.868s (+1.3%)	18.197s (-1.3%)	1.329s	2	1.01x
▲ Vercel	Next.js (Turbopack)	17.826s (+0.9%)	19.468s (+1.2%)	1.641s	2	1.06x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

workflow with 25 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	26.743s (-3.1%)	27.051s (-3.6%)	0.308s	3	1.00x
🌐 Redis	Next.js (Turbopack)	26.760s	27.049s	0.289s	3	1.00x
💻 Local	Next.js (Turbopack)	27.128s	28.053s	0.925s	3	1.01x
🐘 Postgres	Express	27.240s (~)	28.066s (~)	0.826s	3	1.02x
🐘 Postgres	Nitro	27.277s (~)	28.065s (~)	0.788s	3	1.02x
💻 Local	Nitro	27.583s (~)	28.053s (~)	0.470s	3	1.03x
🌐 MongoDB	Next.js (Turbopack)	30.482s	31.045s	0.564s	2	1.14x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	46.028s (+2.9%)	47.737s (+4.3%)	1.709s	2	1.00x
▲ Vercel	Next.js (Turbopack)	50.377s (+6.6% 🔺)	52.253s (+8.0% 🔺)	1.877s	2	1.09x
▲ Vercel	Express	53.606s (+21.0% 🔺)	55.319s (+22.1% 🔺)	1.713s	2	1.16x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

workflow with 50 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	53.541s	54.093s	0.552s	2	1.00x
🐘 Postgres	Express	54.300s (~)	55.094s (~)	0.793s	2	1.01x
🐘 Postgres	Nitro	54.476s (~)	55.105s (~)	0.629s	2	1.02x
💻 Local	Express	54.851s (-3.3%)	55.101s (-3.5%)	0.250s	2	1.02x
💻 Local	Next.js (Turbopack)	55.951s	56.100s	0.149s	2	1.05x
💻 Local	Nitro	56.657s (~)	57.106s (~)	0.450s	2	1.06x
🌐 MongoDB	Next.js (Turbopack)	60.675s	61.065s	0.390s	2	1.13x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	94.771s (-4.4%)	96.975s (-4.2%)	2.204s	1	1.00x
▲ Vercel	Nitro	96.605s (~)	98.384s (~)	1.779s	1	1.02x
▲ Vercel	Next.js (Turbopack)	98.753s (-3.9%)	100.796s (-2.8%)	2.043s	1	1.04x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

Promise.all with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	1.345s	2.006s	0.662s	15	1.00x
🐘 Postgres	Express	1.366s (-4.0%)	2.011s (~)	0.645s	15	1.02x
💻 Local	Express	1.467s (-5.2% 🟢)	2.006s (~)	0.538s	15	1.09x
💻 Local	Nitro	1.498s (-1.1%)	2.005s (~)	0.507s	15	1.11x
💻 Local	Next.js (Turbopack)	1.547s	2.006s	0.459s	15	1.15x
🌐 MongoDB	Next.js (Turbopack)	2.146s	3.009s	0.863s	10	1.60x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-
🐘 Postgres	Nitro	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.312s (-4.2%)	3.407s (-4.8%)	1.095s	9	1.00x
▲ Vercel	Next.js (Turbopack)	2.641s (-1.3%)	4.038s (+7.5% 🔺)	1.397s	8	1.14x
▲ Vercel	Express	2.648s (+5.0%)	4.138s (+13.0% 🔺)	1.490s	8	1.15x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Promise.all with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	2.567s	3.008s	0.440s	10	1.00x
🐘 Postgres	Nitro	2.595s (~)	3.014s (~)	0.419s	10	1.01x
💻 Local	Express	2.600s (-14.1% 🟢)	3.007s (-15.6% 🟢)	0.407s	10	1.01x
🐘 Postgres	Express	2.615s (~)	3.015s (~)	0.399s	10	1.02x
💻 Local	Next.js (Turbopack)	3.041s	3.760s	0.718s	8	1.18x
💻 Local	Nitro	3.121s (+7.5% 🔺)	3.565s (+11.1% 🔺)	0.444s	9	1.22x
🌐 MongoDB	Next.js (Turbopack)	4.747s	5.179s	0.432s	6	1.85x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Next.js (Turbopack)	2.799s (+7.7% 🔺)	4.348s (+21.1% 🔺)	1.550s	8	1.00x
▲ Vercel	Nitro	3.520s (+30.2% 🔺)	4.708s (+23.4% 🔺)	1.188s	7	1.26x
▲ Vercel	Express	3.663s (+47.9% 🔺)	5.183s (+40.8% 🔺)	1.520s	8	1.31x

🔍 Observability: Next.js (Turbopack) | Nitro | Express

Promise.all with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	4.027s (+1.4%)	4.588s (+3.2%)	0.561s	7	1.00x
🐘 Postgres	Nitro	4.032s (+0.7%)	4.590s (+3.1%)	0.558s	7	1.00x
🌐 Redis	Next.js (Turbopack)	4.081s	5.012s	0.931s	6	1.01x
💻 Local	Express	6.787s (-15.7% 🟢)	7.014s (-20.1% 🟢)	0.227s	5	1.69x
💻 Local	Next.js (Turbopack)	7.888s	8.517s	0.629s	4	1.96x
💻 Local	Nitro	8.207s (-1.6%)	8.521s (-5.5% 🟢)	0.314s	4	2.04x
🌐 MongoDB	Next.js (Turbopack)	10.067s	10.684s	0.617s	3	2.50x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Next.js (Turbopack)	5.520s (+44.3% 🔺)	7.146s (+37.1% 🔺)	1.626s	5	1.00x
▲ Vercel	Nitro	7.009s (+153.3% 🔺)	8.352s (+119.6% 🔺)	1.343s	4	1.27x
▲ Vercel	Express	10.371s (+264.0% 🔺)	12.346s (+209.2% 🔺)	1.975s	3	1.88x

🔍 Observability: Next.js (Turbopack) | Nitro | Express

Promise.race with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	1.303s	2.006s	0.703s	15	1.00x
🐘 Postgres	Nitro	1.422s (-1.3%)	2.011s (~)	0.589s	15	1.09x
🐘 Postgres	Express	1.442s (-1.5%)	2.011s (-3.2%)	0.570s	15	1.11x
💻 Local	Express	1.477s (-2.9%)	2.005s (~)	0.528s	15	1.13x
💻 Local	Nitro	1.523s (-1.1%)	2.006s (~)	0.483s	15	1.17x
💻 Local	Next.js (Turbopack)	1.564s	2.073s	0.509s	15	1.20x
🌐 MongoDB	Next.js (Turbopack)	2.174s	3.009s	0.835s	10	1.67x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.131s (-0.9%)	3.290s (-1.5%)	1.159s	10	1.00x
▲ Vercel	Next.js (Turbopack)	2.268s (-27.5% 🟢)	3.823s (-24.1% 🟢)	1.555s	8	1.06x
▲ Vercel	Express	2.591s (+17.2% 🔺)	4.029s (+17.2% 🔺)	1.438s	8	1.22x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Promise.race with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	2.526s (-3.2%)	3.012s (~)	0.486s	10	1.00x
🌐 Redis	Next.js (Turbopack)	2.532s	3.008s	0.477s	10	1.00x
🐘 Postgres	Nitro	2.730s (+7.4% 🔺)	3.015s (~)	0.285s	10	1.08x
💻 Local	Express	2.797s (-9.8% 🟢)	3.108s (-17.3% 🟢)	0.311s	10	1.11x
💻 Local	Next.js (Turbopack)	3.030s	3.760s	0.729s	8	1.20x
💻 Local	Nitro	3.124s (-1.8%)	3.886s (~)	0.762s	8	1.24x
🌐 MongoDB	Next.js (Turbopack)	4.728s	5.177s	0.449s	6	1.87x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.270s (-29.0% 🟢)	3.269s (-37.5% 🟢)	0.998s	10	1.00x
▲ Vercel	Express	2.378s (-32.1% 🟢)	3.827s (-17.2% 🟢)	1.449s	8	1.05x
▲ Vercel	Next.js (Turbopack)	3.093s (-5.5% 🟢)	4.610s (~)	1.517s	7	1.36x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

Promise.race with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	3.975s (-3.5%)	4.590s (~)	0.615s	7	1.00x
🐘 Postgres	Express	4.039s (+2.3%)	4.590s (+3.2%)	0.551s	7	1.02x
🌐 Redis	Next.js (Turbopack)	4.198s	4.725s	0.527s	7	1.06x
💻 Local	Express	7.733s (-14.1% 🟢)	8.267s (-10.8% 🟢)	0.533s	4	1.95x
💻 Local	Next.js (Turbopack)	8.255s	8.766s	0.512s	4	2.08x
💻 Local	Nitro	8.519s (+0.8%)	9.275s (+2.8%)	0.756s	4	2.14x
🌐 MongoDB	Next.js (Turbopack)	10.032s	10.680s	0.648s	3	2.52x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.724s (-27.3% 🟢)	3.647s (-28.7% 🟢)	0.923s	9	1.00x
▲ Vercel	Next.js (Turbopack)	3.571s (-3.3%)	5.369s (+9.7% 🔺)	1.798s	6	1.31x
▲ Vercel	Express	3.789s (+1.4%)	5.100s (-0.7%)	1.311s	6	1.39x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Stream Benchmarks(includes TTFB metrics)

workflow with stream

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	0.138s (-32.2% 🟢)	1.003s (~)	0.009s (-21.6% 🟢)	1.015s (~)	0.877s	10	1.00x
💻 Local	Next.js (Turbopack)	0.170s	1.002s	0.011s	1.017s	0.847s	10	1.24x
🌐 Redis	Next.js (Turbopack)	0.182s	1.000s	0.002s	1.007s	0.825s	10	1.33x
💻 Local	Nitro	0.212s (+7.3% 🔺)	1.003s (~)	0.011s (-2.7%)	1.017s (~)	0.804s	10	1.54x
🐘 Postgres	Express	0.215s (+2.8%)	0.996s (~)	0.002s (+15.4% 🔺)	1.013s (~)	0.797s	10	1.57x
🐘 Postgres	Nitro	0.246s (+9.4% 🔺)	0.993s (~)	0.002s (+41.7% 🔺)	1.014s (~)	0.767s	10	1.79x
🌐 MongoDB	Next.js (Turbopack)	0.474s	0.979s	0.002s	1.009s	0.534s	10	3.45x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	1.606s (+1.9%)	2.405s (-8.7% 🟢)	0.006s (-26.2% 🟢)	3.035s (-2.2%)	1.429s	10	1.00x
▲ Vercel	Nitro	1.658s (+1.2%)	2.506s (-14.3% 🟢)	0.504s (+8896.4% 🔺)	3.490s (+1.5%)	1.833s	10	1.03x
▲ Vercel	Next.js (Turbopack)	1.785s (-4.5%)	3.025s (+6.5% 🔺)	0.011s (+160.5% 🔺)	3.640s (+6.9% 🔺)	1.855s	10	1.11x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

Summary

Fastest Framework by World

Winner determined by most benchmark wins

World	🥇 Fastest Framework	Wins
💻 Local	Express	12/12
🐘 Postgres	Express	7/12
▲ Vercel	Nitro	7/12

Fastest World by Framework

Winner determined by most benchmark wins

Framework	🥇 Fastest World	Wins
Express	💻 Local	6/12
Next.js (Turbopack)	🌐 Redis	9/12
Nitro	🐘 Postgres	6/12

Column Definitions

• Workflow Time: Runtime reported by workflow (completedAt - createdAt) - primary metric
• TTFB: Time to First Byte - time from workflow start until first stream byte received (stream benchmarks only)
• Slurp: Time from first byte to complete stream consumption (stream benchmarks only)
• Wall Time: Total testbench time (trigger workflow + poll for result)
• Overhead: Testbench overhead (Wall Time - Workflow Time)
• Samples: Number of benchmark iterations run
• vs Fastest: How much slower compared to the fastest configuration for this benchmark

Worlds:

• 💻 Local: In-memory filesystem world (local development)
• 🐘 Postgres: PostgreSQL database world (local development)
• ▲ Vercel: Vercel production/preview deployment
• 🌐 Turso: Community world (local development)
• 🌐 MongoDB: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Jazz: Community world (local development)

---

📋 View full workflow run

[pranaygp]

Suggested change

	'Run already finished during setup, skipping',
	'Run already finished during setup, skipping replay',

maybe this would be clearer to the user? (assuming the codepath is only for run replays)

[pranaygp]

or "skipping redundant workflow execution"

[pranaygp]

human: LGTM! great detailed state check and ty for posting that on github PR review comment

[VaguelySerious]

Noticed error logs from 409s falling through the queue, which are confusing for users. Log example:

Queue callback error: Error [WorkflowAPIError]: Cannot update workflow run wrun_01KKVYQ2TNA4PCEH9MA0SHN7DV with status 'completed'. Operation requires status 'running' or 'pending'.
at cP (.next/server/chunks/[root-of-the-server]__fd848f42._.js:63:41)
at async gA (.next/server/chunks/[root-of-the-server]__fd848f42._.js:67:3954)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:86:8)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:60:148157)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:60:146659)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:83:2206)
at async default (.next/server/chunks/[root-of-the-server]__fd848f42._.js:67:7813)
at async u7.processMessage (.next/server/chunks/[root-of-the-server]__fd848f42._.js:62:13992)
at async u7.consume (.next/server/chunks/[root-of-the-server]__fd848f42._.js:62:15197) {
cause: undefined,
status: 409,
code: undefined,
url: 'https://vercel-workflow.com/api/v1/runs/wrun_01KKVYQ2TNA4PCEH9MA0SHN7DV'
}

so I ensures we catch and gracefully return for these cases. I had Claude run an eval on whether this blocks any real use cases. Eval below:

---

Concurrent scenarios and why skipping is safe

Scenario 1: Two handlers race on run_started

Two queue messages for the same run arrive while it's still pending. Both handlers call world.runs.get(), see pending, and try to create a run_started event.

Handler A: runs.get() → pending → events.create(run_started) → ✅ succeeds → continues replay loop
Handler B: runs.get() → pending → events.create(run_started) → ❌ 409 → returns early

Why it's safe: Handler A transitions the run to running and enters the replay loop. The replay loop either completes the workflow or suspends it with queued continuations. Handler B's early exit is a no-op — Handler A guarantees progress.

Location:runtime.ts catch block after run_started event creation.

Scenario 2: Two handlers race on run_completed

Two concurrent replay loops both reach the end of the workflow and try to create run_completed.

Handler A: events.create(run_completed) → ✅ succeeds → returns
Handler B: events.create(run_completed) → ❌ 409 → logs info → returns

Why it's safe: The run is completed. Both handlers return. No continuation is needed.

Location:runtime.ts catch block after run_completed event creation.

Scenario 3: Two handlers race on run_failed

Same as Scenario 2 but for failure. Both handlers detect an error in user code and try to fail the run.

Handler A: events.create(run_failed) → ✅ succeeds → returns
Handler B: events.create(run_failed) → ❌ 409 → logs info → returns

Why it's safe: The run is failed. Both handlers return. No continuation is needed.

Location:runtime.ts catch block after run_failed event creation.

Scenario 4: Run completes while another handler creates hooks

Handler A completes the workflow. Handler B (from an earlier queue message) is still in the suspension handler creating hook events.

Handler A: events.create(run_completed) → ✅ run is now terminal
Handler B: events.create(hook_created) → ❌ 409 (run terminal) → logs info, continues
→ returns from handleSuspension with pendingSteps
→ tries to execute inline step
→ events.create(step_started) → ❌ 410 (run gone) → returns { type: 'gone' }
→ caller returns

Why it's safe: Handler A already completed the workflow. Handler B's hook creation fails, and if any steps were queued, the step executor receives 410 on step_started and exits gracefully. The 410 on step_started is the safety net — even if the suspension handler returns pending steps after a 409, the step executor won't make progress on a finished run.

Location:suspension-handler.ts catch blocks for hook_created and hook_disposed; step-executor.ts 410 handling on step_started.

Scenario 5: Two handlers race on wait_completed

During the replay loop, both handlers see an elapsed wait and try to complete it.

Handler A: events.create(wait_completed) → ✅ succeeds → adds event to cache → continues replay
Handler B: events.create(wait_completed) → ❌ 409 → continue (skip this wait) → continues replay

Why it's safe: Both handlers continue their replay loops. The event log is the same for both (the winning handler's wait_completed is visible to future reads). Subsequent replay from either handler converges to the same state because replay is deterministic and event-sourced.

Location:runtime.ts wait completion loop with continue on 409.

Scenario 6: Two handlers race on step_completed

Two handlers execute the same step concurrently (e.g., both received the step's queue message). Both finish execution and try to create step_completed.

Handler A: events.create(step_completed) → ✅ succeeds → queues workflow continuation → returns
Handler B: events.create(step_completed) → ❌ 409 → returns WITHOUT queuing continuation

Why it's safe: Handler A queues the workflow continuation. Handler B does not — this is correct because only one continuation should be queued per step completion. If both queued, there would be redundant replay (which is safe but wasteful).

Location:step-executor.ts catch on step_completed event creation.

Scenario 7: Two handlers race on step_started

Two handlers try to start the same step. Handler A wins; Handler B gets 409 because the step transitioned to a terminal state.

Handler A: events.create(step_started) → ✅ succeeds → executes step → queues continuation
Handler B: events.create(step_started) → ❌ 409 (step terminal) → returns { type: 'skipped' }
→ caller queues workflow continuation (runtime.ts replay loop)

Why it's safe: Both handlers ensure workflow continuation — Handler A via step completion, Handler B via the replay loop re-queuing the workflow. The redundant replay is safe because event-sourced replay is convergent.

Location:step-executor.ts 409 handling on step_started.

Scenario 8: step_created/wait_created duplicate

During suspension handling, two concurrent replays both try to create the same step or wait event.

Handler A: events.create(step_created) → ✅ succeeds
Handler B: events.create(step_created) → ❌ 409 → logs info → continues

Why it's safe: The step/wait entity already exists. The suspension handler continues processing other items. The step will be executed by whichever handler picks it up from the queue.

Location:suspension-handler.ts catch blocks for step_created and wait_created.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
example-nextjs-workflow-turbopack	Ready	Preview, Comment	Mar 17, 2026 5:00pm
example-nextjs-workflow-webpack	Ready	Preview, Comment	Mar 17, 2026 5:00pm
example-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-astro-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-express-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-fastify-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-hono-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-nitro-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-nuxt-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-sveltekit-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-vite-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workflow-docs	Ready	Preview, Comment, Open in v0	Mar 17, 2026 5:00pm
workflow-nest	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workflow-swc-playground	Ready	Preview, Comment	Mar 17, 2026 5:00pm

[changeset-bot]

🦋 Changeset detected

Latest commit: 9668d09

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 16 packages

Name	Type
@workflow/core	Patch
@workflow/builders	Patch
@workflow/cli	Patch
@workflow/next	Patch
@workflow/nitro	Patch
@workflow/vitest	Patch
@workflow/web-shared	Patch
workflow	Patch
@workflow/world-testing	Patch
@workflow/astro	Patch
@workflow/nest	Patch
@workflow/rollup	Patch
@workflow/sveltekit	Patch
@workflow/vite	Patch
@workflow/nuxt	Patch
@workflow/ai	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

🧪 E2E Test Results

❌ Some tests failed

Summary

	Passed	Failed	Skipped	Total
✅ ▲ Vercel Production	747	0	67	814
✅ 💻 Local Development	770	0	118	888
✅ 📦 Local Production	770	0	118	888
✅ 🐘 Local Postgres	770	0	118	888
✅ 🪟 Windows	71	0	3	74
❌ 🌍 Community Worlds	116	55	15	186
✅ 📋 Other	195	0	27	222
Total	3439	55	466	3960

❌ Failed Tests

🌍 Community Worlds (55 failed)

mongodb (3 failed):

• hookWorkflow is not resumable via public webhook endpoint
• webhookWorkflow
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously

redis (2 failed):

• hookWorkflow is not resumable via public webhook endpoint
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously

turso (50 failed):

• addTenWorkflow
• addTenWorkflow
• wellKnownAgentWorkflow (.well-known/agent)
• should work with react rendering in step
• promiseAllWorkflow
• promiseRaceWorkflow
• promiseAnyWorkflow
• importedStepOnlyWorkflow
• hookWorkflow
• hookWorkflow is not resumable via public webhook endpoint
• webhookWorkflow
• sleepingWorkflow
• parallelSleepWorkflow
• nullByteWorkflow
• workflowAndStepMetadataWorkflow
• fetchWorkflow
• promiseRaceStressTestWorkflow
• error handling error propagation workflow errors nested function calls preserve message and stack trace
• error handling error propagation workflow errors cross-file imports preserve message and stack trace
• error handling error propagation step errors basic step error preserves message and stack trace
• error handling error propagation step errors cross-file step error preserves message and function names in stack
• error handling retry behavior regular Error retries until success
• error handling retry behavior FatalError fails immediately without retries
• error handling retry behavior RetryableError respects custom retryAfter delay
• error handling retry behavior maxRetries=0 disables retries
• error handling catchability FatalError can be caught and detected with FatalError.is()
• hookCleanupTestWorkflow - hook token reuse after workflow completion
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously
• hookDisposeTestWorkflow - hook token reuse after explicit disposal while workflow still running
• stepFunctionPassingWorkflow - step function references can be passed as arguments (without closure vars)
• stepFunctionWithClosureWorkflow - step function with closure variables passed as argument
• closureVariableWorkflow - nested step functions with closure variables
• spawnWorkflowFromStepWorkflow - spawning a child workflow using start() inside a step
• health check (queue-based) - workflow and step endpoints respond to health check messages
• pathsAliasWorkflow - TypeScript path aliases resolve correctly
• Calculator.calculate - static workflow method using static step methods from another class
• AllInOneService.processNumber - static workflow method using sibling static step methods
• ChainableService.processWithThis - static step methods using this to reference the class
• thisSerializationWorkflow - step function invoked with .call() and .apply()
• customSerializationWorkflow - custom class serialization with WORKFLOW_SERIALIZE/WORKFLOW_DESERIALIZE
• instanceMethodStepWorkflow - instance methods with "use step" directive
• crossContextSerdeWorkflow - classes defined in step code are deserializable in workflow context
• stepFunctionAsStartArgWorkflow - step function reference passed as start() argument
• cancelRun - cancelling a running workflow
• cancelRun via CLI - cancelling a running workflow
• pages router addTenWorkflow via pages router
• pages router promiseAllWorkflow via pages router
• pages router sleepingWorkflow via pages router
• hookWithSleepWorkflow - hook payloads delivered correctly with concurrent sleep
• sleepWithSequentialStepsWorkflow - sequential steps work with concurrent sleep (control)

Details by Category

✅ ▲ Vercel Production

App	Passed	Failed	Skipped
✅ astro	67	0	7
✅ example	67	0	7
✅ express	67	0	7
✅ fastify	67	0	7
✅ hono	67	0	7
✅ nextjs-turbopack	72	0	2
✅ nextjs-webpack	72	0	2
✅ nitro	67	0	7
✅ nuxt	67	0	7
✅ sveltekit	67	0	7
✅ vite	67	0	7

✅ 💻 Local Development

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 📦 Local Production

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 🐘 Local Postgres

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 🪟 Windows

App	Passed	Failed	Skipped
✅ nextjs-turbopack	71	0	3

❌ 🌍 Community Worlds

App	Passed	Failed	Skipped
✅ mongodb-dev	3	0	2
❌ mongodb	51	3	3
✅ redis-dev	3	0	2
❌ redis	52	2	3
✅ turso-dev	3	0	2
❌ turso	4	50	3

✅ 📋 Other

App	Passed	Failed	Skipped
✅ e2e-local-dev-nest-stable	65	0	9
✅ e2e-local-postgres-nest-stable	65	0	9
✅ e2e-local-prod-nest-stable	65	0	9

---

📋 View full workflow run

[github-actions]

📊 Benchmark Results

📈 Comparing against baseline from main branch. Green 🟢 = faster, Red 🔺 = slower.

workflow with no steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	0.038s (-15.7% 🟢)	1.005s (~)	0.968s	10	1.00x
💻 Local	Nitro	0.046s (+6.7% 🔺)	1.006s (~)	0.959s	10	1.23x
💻 Local	Next.js (Turbopack)	0.051s	1.006s	0.955s	10	1.36x
🌐 Redis	Next.js (Turbopack)	0.054s	1.005s	0.951s	10	1.44x
🐘 Postgres	Nitro	0.059s (-4.7%)	1.011s (~)	0.952s	10	1.56x
🐘 Postgres	Express	0.059s (-14.8% 🟢)	1.011s (~)	0.952s	10	1.57x
🌐 MongoDB	Next.js (Turbopack)	0.079s	1.008s	0.929s	10	2.11x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	0.465s (-33.5% 🟢)	2.036s (-22.7% 🟢)	1.571s	10	1.00x
▲ Vercel	Express	0.504s (-20.0% 🟢)	2.286s (-14.4% 🟢)	1.781s	10	1.08x
▲ Vercel	Next.js (Turbopack)	0.511s (-17.6% 🟢)	2.559s (+13.2% 🔺)	2.049s	10	1.10x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

workflow with 1 step

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.093s (-3.3%)	2.006s (~)	0.912s	10	1.00x
🌐 Redis	Next.js (Turbopack)	1.122s	2.006s	0.885s	10	1.03x
💻 Local	Nitro	1.130s (~)	2.006s (~)	0.877s	10	1.03x
💻 Local	Next.js (Turbopack)	1.130s	2.006s	0.876s	10	1.03x
🐘 Postgres	Express	1.153s (+0.7%)	2.014s (~)	0.860s	10	1.05x
🐘 Postgres	Nitro	1.158s (+1.1%)	2.012s (~)	0.855s	10	1.06x
🌐 MongoDB	Next.js (Turbopack)	1.307s	2.008s	0.700s	10	1.20x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.040s (-3.0%)	3.224s (-4.4%)	1.184s	10	1.00x
▲ Vercel	Express	2.053s (+0.9%)	3.735s (+12.4% 🔺)	1.682s	10	1.01x
▲ Vercel	Next.js (Turbopack)	2.168s (-2.0%)	3.924s (+13.8% 🔺)	1.756s	10	1.06x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

workflow with 10 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	10.615s (-2.8%)	11.023s (~)	0.408s	3	1.00x
🌐 Redis	Next.js (Turbopack)	10.784s	11.023s	0.239s	3	1.02x
💻 Local	Next.js (Turbopack)	10.812s	11.023s	0.211s	3	1.02x
🐘 Postgres	Nitro	10.927s (~)	11.042s (~)	0.115s	3	1.03x
🐘 Postgres	Express	10.970s (~)	11.375s (~)	0.405s	3	1.03x
💻 Local	Nitro	10.976s (+0.8%)	11.024s (~)	0.048s	3	1.03x
🌐 MongoDB	Next.js (Turbopack)	12.203s	13.015s	0.812s	3	1.15x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	16.779s (-7.4% 🟢)	18.655s (-5.6% 🟢)	1.876s	2	1.00x
▲ Vercel	Nitro	16.868s (+1.3%)	18.197s (-1.3%)	1.329s	2	1.01x
▲ Vercel	Next.js (Turbopack)	17.826s (+0.9%)	19.468s (+1.2%)	1.641s	2	1.06x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

workflow with 25 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	26.743s (-3.1%)	27.051s (-3.6%)	0.308s	3	1.00x
🌐 Redis	Next.js (Turbopack)	26.760s	27.049s	0.289s	3	1.00x
💻 Local	Next.js (Turbopack)	27.128s	28.053s	0.925s	3	1.01x
🐘 Postgres	Express	27.240s (~)	28.066s (~)	0.826s	3	1.02x
🐘 Postgres	Nitro	27.277s (~)	28.065s (~)	0.788s	3	1.02x
💻 Local	Nitro	27.583s (~)	28.053s (~)	0.470s	3	1.03x
🌐 MongoDB	Next.js (Turbopack)	30.482s	31.045s	0.564s	2	1.14x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	46.028s (+2.9%)	47.737s (+4.3%)	1.709s	2	1.00x
▲ Vercel	Next.js (Turbopack)	50.377s (+6.6% 🔺)	52.253s (+8.0% 🔺)	1.877s	2	1.09x
▲ Vercel	Express	53.606s (+21.0% 🔺)	55.319s (+22.1% 🔺)	1.713s	2	1.16x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

workflow with 50 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	53.541s	54.093s	0.552s	2	1.00x
🐘 Postgres	Express	54.300s (~)	55.094s (~)	0.793s	2	1.01x
🐘 Postgres	Nitro	54.476s (~)	55.105s (~)	0.629s	2	1.02x
💻 Local	Express	54.851s (-3.3%)	55.101s (-3.5%)	0.250s	2	1.02x
💻 Local	Next.js (Turbopack)	55.951s	56.100s	0.149s	2	1.05x
💻 Local	Nitro	56.657s (~)	57.106s (~)	0.450s	2	1.06x
🌐 MongoDB	Next.js (Turbopack)	60.675s	61.065s	0.390s	2	1.13x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	94.771s (-4.4%)	96.975s (-4.2%)	2.204s	1	1.00x
▲ Vercel	Nitro	96.605s (~)	98.384s (~)	1.779s	1	1.02x
▲ Vercel	Next.js (Turbopack)	98.753s (-3.9%)	100.796s (-2.8%)	2.043s	1	1.04x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

Promise.all with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	1.345s	2.006s	0.662s	15	1.00x
🐘 Postgres	Express	1.366s (-4.0%)	2.011s (~)	0.645s	15	1.02x
💻 Local	Express	1.467s (-5.2% 🟢)	2.006s (~)	0.538s	15	1.09x
💻 Local	Nitro	1.498s (-1.1%)	2.005s (~)	0.507s	15	1.11x
💻 Local	Next.js (Turbopack)	1.547s	2.006s	0.459s	15	1.15x
🌐 MongoDB	Next.js (Turbopack)	2.146s	3.009s	0.863s	10	1.60x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-
🐘 Postgres	Nitro	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.312s (-4.2%)	3.407s (-4.8%)	1.095s	9	1.00x
▲ Vercel	Next.js (Turbopack)	2.641s (-1.3%)	4.038s (+7.5% 🔺)	1.397s	8	1.14x
▲ Vercel	Express	2.648s (+5.0%)	4.138s (+13.0% 🔺)	1.490s	8	1.15x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Promise.all with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	2.567s	3.008s	0.440s	10	1.00x
🐘 Postgres	Nitro	2.595s (~)	3.014s (~)	0.419s	10	1.01x
💻 Local	Express	2.600s (-14.1% 🟢)	3.007s (-15.6% 🟢)	0.407s	10	1.01x
🐘 Postgres	Express	2.615s (~)	3.015s (~)	0.399s	10	1.02x
💻 Local	Next.js (Turbopack)	3.041s	3.760s	0.718s	8	1.18x
💻 Local	Nitro	3.121s (+7.5% 🔺)	3.565s (+11.1% 🔺)	0.444s	9	1.22x
🌐 MongoDB	Next.js (Turbopack)	4.747s	5.179s	0.432s	6	1.85x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Next.js (Turbopack)	2.799s (+7.7% 🔺)	4.348s (+21.1% 🔺)	1.550s	8	1.00x
▲ Vercel	Nitro	3.520s (+30.2% 🔺)	4.708s (+23.4% 🔺)	1.188s	7	1.26x
▲ Vercel	Express	3.663s (+47.9% 🔺)	5.183s (+40.8% 🔺)	1.520s	8	1.31x

🔍 Observability: Next.js (Turbopack) | Nitro | Express

Promise.all with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	4.027s (+1.4%)	4.588s (+3.2%)	0.561s	7	1.00x
🐘 Postgres	Nitro	4.032s (+0.7%)	4.590s (+3.1%)	0.558s	7	1.00x
🌐 Redis	Next.js (Turbopack)	4.081s	5.012s	0.931s	6	1.01x
💻 Local	Express	6.787s (-15.7% 🟢)	7.014s (-20.1% 🟢)	0.227s	5	1.69x
💻 Local	Next.js (Turbopack)	7.888s	8.517s	0.629s	4	1.96x
💻 Local	Nitro	8.207s (-1.6%)	8.521s (-5.5% 🟢)	0.314s	4	2.04x
🌐 MongoDB	Next.js (Turbopack)	10.067s	10.684s	0.617s	3	2.50x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Next.js (Turbopack)	5.520s (+44.3% 🔺)	7.146s (+37.1% 🔺)	1.626s	5	1.00x
▲ Vercel	Nitro	7.009s (+153.3% 🔺)	8.352s (+119.6% 🔺)	1.343s	4	1.27x
▲ Vercel	Express	10.371s (+264.0% 🔺)	12.346s (+209.2% 🔺)	1.975s	3	1.88x

🔍 Observability: Next.js (Turbopack) | Nitro | Express

Promise.race with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	1.303s	2.006s	0.703s	15	1.00x
🐘 Postgres	Nitro	1.422s (-1.3%)	2.011s (~)	0.589s	15	1.09x
🐘 Postgres	Express	1.442s (-1.5%)	2.011s (-3.2%)	0.570s	15	1.11x
💻 Local	Express	1.477s (-2.9%)	2.005s (~)	0.528s	15	1.13x
💻 Local	Nitro	1.523s (-1.1%)	2.006s (~)	0.483s	15	1.17x
💻 Local	Next.js (Turbopack)	1.564s	2.073s	0.509s	15	1.20x
🌐 MongoDB	Next.js (Turbopack)	2.174s	3.009s	0.835s	10	1.67x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.131s (-0.9%)	3.290s (-1.5%)	1.159s	10	1.00x
▲ Vercel	Next.js (Turbopack)	2.268s (-27.5% 🟢)	3.823s (-24.1% 🟢)	1.555s	8	1.06x
▲ Vercel	Express	2.591s (+17.2% 🔺)	4.029s (+17.2% 🔺)	1.438s	8	1.22x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Promise.race with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	2.526s (-3.2%)	3.012s (~)	0.486s	10	1.00x
🌐 Redis	Next.js (Turbopack)	2.532s	3.008s	0.477s	10	1.00x
🐘 Postgres	Nitro	2.730s (+7.4% 🔺)	3.015s (~)	0.285s	10	1.08x
💻 Local	Express	2.797s (-9.8% 🟢)	3.108s (-17.3% 🟢)	0.311s	10	1.11x
💻 Local	Next.js (Turbopack)	3.030s	3.760s	0.729s	8	1.20x
💻 Local	Nitro	3.124s (-1.8%)	3.886s (~)	0.762s	8	1.24x
🌐 MongoDB	Next.js (Turbopack)	4.728s	5.177s	0.449s	6	1.87x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.270s (-29.0% 🟢)	3.269s (-37.5% 🟢)	0.998s	10	1.00x
▲ Vercel	Express	2.378s (-32.1% 🟢)	3.827s (-17.2% 🟢)	1.449s	8	1.05x
▲ Vercel	Next.js (Turbopack)	3.093s (-5.5% 🟢)	4.610s (~)	1.517s	7	1.36x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

Promise.race with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	3.975s (-3.5%)	4.590s (~)	0.615s	7	1.00x
🐘 Postgres	Express	4.039s (+2.3%)	4.590s (+3.2%)	0.551s	7	1.02x
🌐 Redis	Next.js (Turbopack)	4.198s	4.725s	0.527s	7	1.06x
💻 Local	Express	7.733s (-14.1% 🟢)	8.267s (-10.8% 🟢)	0.533s	4	1.95x
💻 Local	Next.js (Turbopack)	8.255s	8.766s	0.512s	4	2.08x
💻 Local	Nitro	8.519s (+0.8%)	9.275s (+2.8%)	0.756s	4	2.14x
🌐 MongoDB	Next.js (Turbopack)	10.032s	10.680s	0.648s	3	2.52x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.724s (-27.3% 🟢)	3.647s (-28.7% 🟢)	0.923s	9	1.00x
▲ Vercel	Next.js (Turbopack)	3.571s (-3.3%)	5.369s (+9.7% 🔺)	1.798s	6	1.31x
▲ Vercel	Express	3.789s (+1.4%)	5.100s (-0.7%)	1.311s	6	1.39x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Stream Benchmarks(includes TTFB metrics)

workflow with stream

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	0.138s (-32.2% 🟢)	1.003s (~)	0.009s (-21.6% 🟢)	1.015s (~)	0.877s	10	1.00x
💻 Local	Next.js (Turbopack)	0.170s	1.002s	0.011s	1.017s	0.847s	10	1.24x
🌐 Redis	Next.js (Turbopack)	0.182s	1.000s	0.002s	1.007s	0.825s	10	1.33x
💻 Local	Nitro	0.212s (+7.3% 🔺)	1.003s (~)	0.011s (-2.7%)	1.017s (~)	0.804s	10	1.54x
🐘 Postgres	Express	0.215s (+2.8%)	0.996s (~)	0.002s (+15.4% 🔺)	1.013s (~)	0.797s	10	1.57x
🐘 Postgres	Nitro	0.246s (+9.4% 🔺)	0.993s (~)	0.002s (+41.7% 🔺)	1.014s (~)	0.767s	10	1.79x
🌐 MongoDB	Next.js (Turbopack)	0.474s	0.979s	0.002s	1.009s	0.534s	10	3.45x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	1.606s (+1.9%)	2.405s (-8.7% 🟢)	0.006s (-26.2% 🟢)	3.035s (-2.2%)	1.429s	10	1.00x
▲ Vercel	Nitro	1.658s (+1.2%)	2.506s (-14.3% 🟢)	0.504s (+8896.4% 🔺)	3.490s (+1.5%)	1.833s	10	1.03x
▲ Vercel	Next.js (Turbopack)	1.785s (-4.5%)	3.025s (+6.5% 🔺)	0.011s (+160.5% 🔺)	3.640s (+6.9% 🔺)	1.855s	10	1.11x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

Summary

Fastest Framework by World

Winner determined by most benchmark wins

World	🥇 Fastest Framework	Wins
💻 Local	Express	12/12
🐘 Postgres	Express	7/12
▲ Vercel	Nitro	7/12

Fastest World by Framework

Winner determined by most benchmark wins

Framework	🥇 Fastest World	Wins
Express	💻 Local	6/12
Next.js (Turbopack)	🌐 Redis	9/12
Nitro	🐘 Postgres	6/12

Column Definitions

• Workflow Time: Runtime reported by workflow (completedAt - createdAt) - primary metric
• TTFB: Time to First Byte - time from workflow start until first stream byte received (stream benchmarks only)
• Slurp: Time from first byte to complete stream consumption (stream benchmarks only)
• Wall Time: Total testbench time (trigger workflow + poll for result)
• Overhead: Testbench overhead (Wall Time - Workflow Time)
• Samples: Number of benchmark iterations run
• vs Fastest: How much slower compared to the fastest configuration for this benchmark

Worlds:

• 💻 Local: In-memory filesystem world (local development)
• 🐘 Postgres: PostgreSQL database world (local development)
• ▲ Vercel: Vercel production/preview deployment
• 🌐 Turso: Community world (local development)
• 🌐 MongoDB: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Jazz: Community world (local development)

---

📋 View full workflow run

[pranaygp]

Suggested change

	'Run already finished during setup, skipping',
	'Run already finished during setup, skipping replay',

maybe this would be clearer to the user? (assuming the codepath is only for run replays)

[pranaygp]

or "skipping redundant workflow execution"

[pranaygp]

human: LGTM! great detailed state check and ty for posting that on github PR review comment

[VaguelySerious]

Noticed error logs from 409s falling through the queue, which are confusing for users. Log example:

Queue callback error: Error [WorkflowAPIError]: Cannot update workflow run wrun_01KKVYQ2TNA4PCEH9MA0SHN7DV with status 'completed'. Operation requires status 'running' or 'pending'.
at cP (.next/server/chunks/[root-of-the-server]__fd848f42._.js:63:41)
at async gA (.next/server/chunks/[root-of-the-server]__fd848f42._.js:67:3954)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:86:8)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:60:148157)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:60:146659)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:83:2206)
at async default (.next/server/chunks/[root-of-the-server]__fd848f42._.js:67:7813)
at async u7.processMessage (.next/server/chunks/[root-of-the-server]__fd848f42._.js:62:13992)
at async u7.consume (.next/server/chunks/[root-of-the-server]__fd848f42._.js:62:15197) {
cause: undefined,
status: 409,
code: undefined,
url: 'https://vercel-workflow.com/api/v1/runs/wrun_01KKVYQ2TNA4PCEH9MA0SHN7DV'
}

so I ensures we catch and gracefully return for these cases. I had Claude run an eval on whether this blocks any real use cases. Eval below:

---

Concurrent scenarios and why skipping is safe

Scenario 1: Two handlers race on run_started

Two queue messages for the same run arrive while it's still pending. Both handlers call world.runs.get(), see pending, and try to create a run_started event.

Handler A: runs.get() → pending → events.create(run_started) → ✅ succeeds → continues replay loop
Handler B: runs.get() → pending → events.create(run_started) → ❌ 409 → returns early

Why it's safe: Handler A transitions the run to running and enters the replay loop. The replay loop either completes the workflow or suspends it with queued continuations. Handler B's early exit is a no-op — Handler A guarantees progress.

Location:runtime.ts catch block after run_started event creation.

Scenario 2: Two handlers race on run_completed

Two concurrent replay loops both reach the end of the workflow and try to create run_completed.

Handler A: events.create(run_completed) → ✅ succeeds → returns
Handler B: events.create(run_completed) → ❌ 409 → logs info → returns

Why it's safe: The run is completed. Both handlers return. No continuation is needed.

Location:runtime.ts catch block after run_completed event creation.

Scenario 3: Two handlers race on run_failed

Same as Scenario 2 but for failure. Both handlers detect an error in user code and try to fail the run.

Handler A: events.create(run_failed) → ✅ succeeds → returns
Handler B: events.create(run_failed) → ❌ 409 → logs info → returns

Why it's safe: The run is failed. Both handlers return. No continuation is needed.

Location:runtime.ts catch block after run_failed event creation.

Scenario 4: Run completes while another handler creates hooks

Handler A completes the workflow. Handler B (from an earlier queue message) is still in the suspension handler creating hook events.

Handler A: events.create(run_completed) → ✅ run is now terminal
Handler B: events.create(hook_created) → ❌ 409 (run terminal) → logs info, continues
→ returns from handleSuspension with pendingSteps
→ tries to execute inline step
→ events.create(step_started) → ❌ 410 (run gone) → returns { type: 'gone' }
→ caller returns

Why it's safe: Handler A already completed the workflow. Handler B's hook creation fails, and if any steps were queued, the step executor receives 410 on step_started and exits gracefully. The 410 on step_started is the safety net — even if the suspension handler returns pending steps after a 409, the step executor won't make progress on a finished run.

Location:suspension-handler.ts catch blocks for hook_created and hook_disposed; step-executor.ts 410 handling on step_started.

Scenario 5: Two handlers race on wait_completed

During the replay loop, both handlers see an elapsed wait and try to complete it.

Handler A: events.create(wait_completed) → ✅ succeeds → adds event to cache → continues replay
Handler B: events.create(wait_completed) → ❌ 409 → continue (skip this wait) → continues replay

Why it's safe: Both handlers continue their replay loops. The event log is the same for both (the winning handler's wait_completed is visible to future reads). Subsequent replay from either handler converges to the same state because replay is deterministic and event-sourced.

Location:runtime.ts wait completion loop with continue on 409.

Scenario 6: Two handlers race on step_completed

Two handlers execute the same step concurrently (e.g., both received the step's queue message). Both finish execution and try to create step_completed.

Handler A: events.create(step_completed) → ✅ succeeds → queues workflow continuation → returns
Handler B: events.create(step_completed) → ❌ 409 → returns WITHOUT queuing continuation

Why it's safe: Handler A queues the workflow continuation. Handler B does not — this is correct because only one continuation should be queued per step completion. If both queued, there would be redundant replay (which is safe but wasteful).

Location:step-executor.ts catch on step_completed event creation.

Scenario 7: Two handlers race on step_started

Two handlers try to start the same step. Handler A wins; Handler B gets 409 because the step transitioned to a terminal state.

Handler A: events.create(step_started) → ✅ succeeds → executes step → queues continuation
Handler B: events.create(step_started) → ❌ 409 (step terminal) → returns { type: 'skipped' }
→ caller queues workflow continuation (runtime.ts replay loop)

Why it's safe: Both handlers ensure workflow continuation — Handler A via step completion, Handler B via the replay loop re-queuing the workflow. The redundant replay is safe because event-sourced replay is convergent.

Location:step-executor.ts 409 handling on step_started.

Scenario 8: step_created/wait_created duplicate

During suspension handling, two concurrent replays both try to create the same step or wait event.

Handler A: events.create(step_created) → ✅ succeeds
Handler B: events.create(step_created) → ❌ 409 → logs info → continues

Why it's safe: The step/wait entity already exists. The suspension handler continues processing other items. The step will be executed by whichever handler picks it up from the queue.

Location:suspension-handler.ts catch blocks for step_created and wait_created.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
example-nextjs-workflow-turbopack	Ready	Preview, Comment	Mar 17, 2026 5:00pm
example-nextjs-workflow-webpack	Ready	Preview, Comment	Mar 17, 2026 5:00pm
example-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-astro-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-express-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-fastify-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-hono-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-nitro-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-nuxt-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-sveltekit-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-vite-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workflow-docs	Ready	Preview, Comment, Open in v0	Mar 17, 2026 5:00pm
workflow-nest	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workflow-swc-playground	Ready	Preview, Comment	Mar 17, 2026 5:00pm

[changeset-bot]

🦋 Changeset detected

Latest commit: 9668d09

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 16 packages

Name	Type
@workflow/core	Patch
@workflow/builders	Patch
@workflow/cli	Patch
@workflow/next	Patch
@workflow/nitro	Patch
@workflow/vitest	Patch
@workflow/web-shared	Patch
workflow	Patch
@workflow/world-testing	Patch
@workflow/astro	Patch
@workflow/nest	Patch
@workflow/rollup	Patch
@workflow/sveltekit	Patch
@workflow/vite	Patch
@workflow/nuxt	Patch
@workflow/ai	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

🧪 E2E Test Results

❌ Some tests failed

Summary

	Passed	Failed	Skipped	Total
✅ ▲ Vercel Production	747	0	67	814
✅ 💻 Local Development	770	0	118	888
✅ 📦 Local Production	770	0	118	888
✅ 🐘 Local Postgres	770	0	118	888
✅ 🪟 Windows	71	0	3	74
❌ 🌍 Community Worlds	116	55	15	186
✅ 📋 Other	195	0	27	222
Total	3439	55	466	3960

❌ Failed Tests

🌍 Community Worlds (55 failed)

mongodb (3 failed):

• hookWorkflow is not resumable via public webhook endpoint
• webhookWorkflow
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously

redis (2 failed):

• hookWorkflow is not resumable via public webhook endpoint
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously

turso (50 failed):

• addTenWorkflow
• addTenWorkflow
• wellKnownAgentWorkflow (.well-known/agent)
• should work with react rendering in step
• promiseAllWorkflow
• promiseRaceWorkflow
• promiseAnyWorkflow
• importedStepOnlyWorkflow
• hookWorkflow
• hookWorkflow is not resumable via public webhook endpoint
• webhookWorkflow
• sleepingWorkflow
• parallelSleepWorkflow
• nullByteWorkflow
• workflowAndStepMetadataWorkflow
• fetchWorkflow
• promiseRaceStressTestWorkflow
• error handling error propagation workflow errors nested function calls preserve message and stack trace
• error handling error propagation workflow errors cross-file imports preserve message and stack trace
• error handling error propagation step errors basic step error preserves message and stack trace
• error handling error propagation step errors cross-file step error preserves message and function names in stack
• error handling retry behavior regular Error retries until success
• error handling retry behavior FatalError fails immediately without retries
• error handling retry behavior RetryableError respects custom retryAfter delay
• error handling retry behavior maxRetries=0 disables retries
• error handling catchability FatalError can be caught and detected with FatalError.is()
• hookCleanupTestWorkflow - hook token reuse after workflow completion
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously
• hookDisposeTestWorkflow - hook token reuse after explicit disposal while workflow still running
• stepFunctionPassingWorkflow - step function references can be passed as arguments (without closure vars)
• stepFunctionWithClosureWorkflow - step function with closure variables passed as argument
• closureVariableWorkflow - nested step functions with closure variables
• spawnWorkflowFromStepWorkflow - spawning a child workflow using start() inside a step
• health check (queue-based) - workflow and step endpoints respond to health check messages
• pathsAliasWorkflow - TypeScript path aliases resolve correctly
• Calculator.calculate - static workflow method using static step methods from another class
• AllInOneService.processNumber - static workflow method using sibling static step methods
• ChainableService.processWithThis - static step methods using this to reference the class
• thisSerializationWorkflow - step function invoked with .call() and .apply()
• customSerializationWorkflow - custom class serialization with WORKFLOW_SERIALIZE/WORKFLOW_DESERIALIZE
• instanceMethodStepWorkflow - instance methods with "use step" directive
• crossContextSerdeWorkflow - classes defined in step code are deserializable in workflow context
• stepFunctionAsStartArgWorkflow - step function reference passed as start() argument
• cancelRun - cancelling a running workflow
• cancelRun via CLI - cancelling a running workflow
• pages router addTenWorkflow via pages router
• pages router promiseAllWorkflow via pages router
• pages router sleepingWorkflow via pages router
• hookWithSleepWorkflow - hook payloads delivered correctly with concurrent sleep
• sleepWithSequentialStepsWorkflow - sequential steps work with concurrent sleep (control)

Details by Category

✅ ▲ Vercel Production

App	Passed	Failed	Skipped
✅ astro	67	0	7
✅ example	67	0	7
✅ express	67	0	7
✅ fastify	67	0	7
✅ hono	67	0	7
✅ nextjs-turbopack	72	0	2
✅ nextjs-webpack	72	0	2
✅ nitro	67	0	7
✅ nuxt	67	0	7
✅ sveltekit	67	0	7
✅ vite	67	0	7

✅ 💻 Local Development

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 📦 Local Production

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 🐘 Local Postgres

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 🪟 Windows

App	Passed	Failed	Skipped
✅ nextjs-turbopack	71	0	3

❌ 🌍 Community Worlds

App	Passed	Failed	Skipped
✅ mongodb-dev	3	0	2
❌ mongodb	51	3	3
✅ redis-dev	3	0	2
❌ redis	52	2	3
✅ turso-dev	3	0	2
❌ turso	4	50	3

✅ 📋 Other

App	Passed	Failed	Skipped
✅ e2e-local-dev-nest-stable	65	0	9
✅ e2e-local-postgres-nest-stable	65	0	9
✅ e2e-local-prod-nest-stable	65	0	9

---

📋 View full workflow run

[github-actions]

📊 Benchmark Results

📈 Comparing against baseline from main branch. Green 🟢 = faster, Red 🔺 = slower.

workflow with no steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	0.038s (-15.7% 🟢)	1.005s (~)	0.968s	10	1.00x
💻 Local	Nitro	0.046s (+6.7% 🔺)	1.006s (~)	0.959s	10	1.23x
💻 Local	Next.js (Turbopack)	0.051s	1.006s	0.955s	10	1.36x
🌐 Redis	Next.js (Turbopack)	0.054s	1.005s	0.951s	10	1.44x
🐘 Postgres	Nitro	0.059s (-4.7%)	1.011s (~)	0.952s	10	1.56x
🐘 Postgres	Express	0.059s (-14.8% 🟢)	1.011s (~)	0.952s	10	1.57x
🌐 MongoDB	Next.js (Turbopack)	0.079s	1.008s	0.929s	10	2.11x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	0.465s (-33.5% 🟢)	2.036s (-22.7% 🟢)	1.571s	10	1.00x
▲ Vercel	Express	0.504s (-20.0% 🟢)	2.286s (-14.4% 🟢)	1.781s	10	1.08x
▲ Vercel	Next.js (Turbopack)	0.511s (-17.6% 🟢)	2.559s (+13.2% 🔺)	2.049s	10	1.10x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

workflow with 1 step

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.093s (-3.3%)	2.006s (~)	0.912s	10	1.00x
🌐 Redis	Next.js (Turbopack)	1.122s	2.006s	0.885s	10	1.03x
💻 Local	Nitro	1.130s (~)	2.006s (~)	0.877s	10	1.03x
💻 Local	Next.js (Turbopack)	1.130s	2.006s	0.876s	10	1.03x
🐘 Postgres	Express	1.153s (+0.7%)	2.014s (~)	0.860s	10	1.05x
🐘 Postgres	Nitro	1.158s (+1.1%)	2.012s (~)	0.855s	10	1.06x
🌐 MongoDB	Next.js (Turbopack)	1.307s	2.008s	0.700s	10	1.20x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.040s (-3.0%)	3.224s (-4.4%)	1.184s	10	1.00x
▲ Vercel	Express	2.053s (+0.9%)	3.735s (+12.4% 🔺)	1.682s	10	1.01x
▲ Vercel	Next.js (Turbopack)	2.168s (-2.0%)	3.924s (+13.8% 🔺)	1.756s	10	1.06x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

workflow with 10 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	10.615s (-2.8%)	11.023s (~)	0.408s	3	1.00x
🌐 Redis	Next.js (Turbopack)	10.784s	11.023s	0.239s	3	1.02x
💻 Local	Next.js (Turbopack)	10.812s	11.023s	0.211s	3	1.02x
🐘 Postgres	Nitro	10.927s (~)	11.042s (~)	0.115s	3	1.03x
🐘 Postgres	Express	10.970s (~)	11.375s (~)	0.405s	3	1.03x
💻 Local	Nitro	10.976s (+0.8%)	11.024s (~)	0.048s	3	1.03x
🌐 MongoDB	Next.js (Turbopack)	12.203s	13.015s	0.812s	3	1.15x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	16.779s (-7.4% 🟢)	18.655s (-5.6% 🟢)	1.876s	2	1.00x
▲ Vercel	Nitro	16.868s (+1.3%)	18.197s (-1.3%)	1.329s	2	1.01x
▲ Vercel	Next.js (Turbopack)	17.826s (+0.9%)	19.468s (+1.2%)	1.641s	2	1.06x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

workflow with 25 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	26.743s (-3.1%)	27.051s (-3.6%)	0.308s	3	1.00x
🌐 Redis	Next.js (Turbopack)	26.760s	27.049s	0.289s	3	1.00x
💻 Local	Next.js (Turbopack)	27.128s	28.053s	0.925s	3	1.01x
🐘 Postgres	Express	27.240s (~)	28.066s (~)	0.826s	3	1.02x
🐘 Postgres	Nitro	27.277s (~)	28.065s (~)	0.788s	3	1.02x
💻 Local	Nitro	27.583s (~)	28.053s (~)	0.470s	3	1.03x
🌐 MongoDB	Next.js (Turbopack)	30.482s	31.045s	0.564s	2	1.14x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	46.028s (+2.9%)	47.737s (+4.3%)	1.709s	2	1.00x
▲ Vercel	Next.js (Turbopack)	50.377s (+6.6% 🔺)	52.253s (+8.0% 🔺)	1.877s	2	1.09x
▲ Vercel	Express	53.606s (+21.0% 🔺)	55.319s (+22.1% 🔺)	1.713s	2	1.16x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

workflow with 50 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	53.541s	54.093s	0.552s	2	1.00x
🐘 Postgres	Express	54.300s (~)	55.094s (~)	0.793s	2	1.01x
🐘 Postgres	Nitro	54.476s (~)	55.105s (~)	0.629s	2	1.02x
💻 Local	Express	54.851s (-3.3%)	55.101s (-3.5%)	0.250s	2	1.02x
💻 Local	Next.js (Turbopack)	55.951s	56.100s	0.149s	2	1.05x
💻 Local	Nitro	56.657s (~)	57.106s (~)	0.450s	2	1.06x
🌐 MongoDB	Next.js (Turbopack)	60.675s	61.065s	0.390s	2	1.13x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	94.771s (-4.4%)	96.975s (-4.2%)	2.204s	1	1.00x
▲ Vercel	Nitro	96.605s (~)	98.384s (~)	1.779s	1	1.02x
▲ Vercel	Next.js (Turbopack)	98.753s (-3.9%)	100.796s (-2.8%)	2.043s	1	1.04x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

Promise.all with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	1.345s	2.006s	0.662s	15	1.00x
🐘 Postgres	Express	1.366s (-4.0%)	2.011s (~)	0.645s	15	1.02x
💻 Local	Express	1.467s (-5.2% 🟢)	2.006s (~)	0.538s	15	1.09x
💻 Local	Nitro	1.498s (-1.1%)	2.005s (~)	0.507s	15	1.11x
💻 Local	Next.js (Turbopack)	1.547s	2.006s	0.459s	15	1.15x
🌐 MongoDB	Next.js (Turbopack)	2.146s	3.009s	0.863s	10	1.60x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-
🐘 Postgres	Nitro	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.312s (-4.2%)	3.407s (-4.8%)	1.095s	9	1.00x
▲ Vercel	Next.js (Turbopack)	2.641s (-1.3%)	4.038s (+7.5% 🔺)	1.397s	8	1.14x
▲ Vercel	Express	2.648s (+5.0%)	4.138s (+13.0% 🔺)	1.490s	8	1.15x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Promise.all with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	2.567s	3.008s	0.440s	10	1.00x
🐘 Postgres	Nitro	2.595s (~)	3.014s (~)	0.419s	10	1.01x
💻 Local	Express	2.600s (-14.1% 🟢)	3.007s (-15.6% 🟢)	0.407s	10	1.01x
🐘 Postgres	Express	2.615s (~)	3.015s (~)	0.399s	10	1.02x
💻 Local	Next.js (Turbopack)	3.041s	3.760s	0.718s	8	1.18x
💻 Local	Nitro	3.121s (+7.5% 🔺)	3.565s (+11.1% 🔺)	0.444s	9	1.22x
🌐 MongoDB	Next.js (Turbopack)	4.747s	5.179s	0.432s	6	1.85x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Next.js (Turbopack)	2.799s (+7.7% 🔺)	4.348s (+21.1% 🔺)	1.550s	8	1.00x
▲ Vercel	Nitro	3.520s (+30.2% 🔺)	4.708s (+23.4% 🔺)	1.188s	7	1.26x
▲ Vercel	Express	3.663s (+47.9% 🔺)	5.183s (+40.8% 🔺)	1.520s	8	1.31x

🔍 Observability: Next.js (Turbopack) | Nitro | Express

Promise.all with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	4.027s (+1.4%)	4.588s (+3.2%)	0.561s	7	1.00x
🐘 Postgres	Nitro	4.032s (+0.7%)	4.590s (+3.1%)	0.558s	7	1.00x
🌐 Redis	Next.js (Turbopack)	4.081s	5.012s	0.931s	6	1.01x
💻 Local	Express	6.787s (-15.7% 🟢)	7.014s (-20.1% 🟢)	0.227s	5	1.69x
💻 Local	Next.js (Turbopack)	7.888s	8.517s	0.629s	4	1.96x
💻 Local	Nitro	8.207s (-1.6%)	8.521s (-5.5% 🟢)	0.314s	4	2.04x
🌐 MongoDB	Next.js (Turbopack)	10.067s	10.684s	0.617s	3	2.50x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Next.js (Turbopack)	5.520s (+44.3% 🔺)	7.146s (+37.1% 🔺)	1.626s	5	1.00x
▲ Vercel	Nitro	7.009s (+153.3% 🔺)	8.352s (+119.6% 🔺)	1.343s	4	1.27x
▲ Vercel	Express	10.371s (+264.0% 🔺)	12.346s (+209.2% 🔺)	1.975s	3	1.88x

🔍 Observability: Next.js (Turbopack) | Nitro | Express

Promise.race with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	1.303s	2.006s	0.703s	15	1.00x
🐘 Postgres	Nitro	1.422s (-1.3%)	2.011s (~)	0.589s	15	1.09x
🐘 Postgres	Express	1.442s (-1.5%)	2.011s (-3.2%)	0.570s	15	1.11x
💻 Local	Express	1.477s (-2.9%)	2.005s (~)	0.528s	15	1.13x
💻 Local	Nitro	1.523s (-1.1%)	2.006s (~)	0.483s	15	1.17x
💻 Local	Next.js (Turbopack)	1.564s	2.073s	0.509s	15	1.20x
🌐 MongoDB	Next.js (Turbopack)	2.174s	3.009s	0.835s	10	1.67x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.131s (-0.9%)	3.290s (-1.5%)	1.159s	10	1.00x
▲ Vercel	Next.js (Turbopack)	2.268s (-27.5% 🟢)	3.823s (-24.1% 🟢)	1.555s	8	1.06x
▲ Vercel	Express	2.591s (+17.2% 🔺)	4.029s (+17.2% 🔺)	1.438s	8	1.22x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Promise.race with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	2.526s (-3.2%)	3.012s (~)	0.486s	10	1.00x
🌐 Redis	Next.js (Turbopack)	2.532s	3.008s	0.477s	10	1.00x
🐘 Postgres	Nitro	2.730s (+7.4% 🔺)	3.015s (~)	0.285s	10	1.08x
💻 Local	Express	2.797s (-9.8% 🟢)	3.108s (-17.3% 🟢)	0.311s	10	1.11x
💻 Local	Next.js (Turbopack)	3.030s	3.760s	0.729s	8	1.20x
💻 Local	Nitro	3.124s (-1.8%)	3.886s (~)	0.762s	8	1.24x
🌐 MongoDB	Next.js (Turbopack)	4.728s	5.177s	0.449s	6	1.87x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.270s (-29.0% 🟢)	3.269s (-37.5% 🟢)	0.998s	10	1.00x
▲ Vercel	Express	2.378s (-32.1% 🟢)	3.827s (-17.2% 🟢)	1.449s	8	1.05x
▲ Vercel	Next.js (Turbopack)	3.093s (-5.5% 🟢)	4.610s (~)	1.517s	7	1.36x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

Promise.race with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	3.975s (-3.5%)	4.590s (~)	0.615s	7	1.00x
🐘 Postgres	Express	4.039s (+2.3%)	4.590s (+3.2%)	0.551s	7	1.02x
🌐 Redis	Next.js (Turbopack)	4.198s	4.725s	0.527s	7	1.06x
💻 Local	Express	7.733s (-14.1% 🟢)	8.267s (-10.8% 🟢)	0.533s	4	1.95x
💻 Local	Next.js (Turbopack)	8.255s	8.766s	0.512s	4	2.08x
💻 Local	Nitro	8.519s (+0.8%)	9.275s (+2.8%)	0.756s	4	2.14x
🌐 MongoDB	Next.js (Turbopack)	10.032s	10.680s	0.648s	3	2.52x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.724s (-27.3% 🟢)	3.647s (-28.7% 🟢)	0.923s	9	1.00x
▲ Vercel	Next.js (Turbopack)	3.571s (-3.3%)	5.369s (+9.7% 🔺)	1.798s	6	1.31x
▲ Vercel	Express	3.789s (+1.4%)	5.100s (-0.7%)	1.311s	6	1.39x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Stream Benchmarks(includes TTFB metrics)

workflow with stream

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	0.138s (-32.2% 🟢)	1.003s (~)	0.009s (-21.6% 🟢)	1.015s (~)	0.877s	10	1.00x
💻 Local	Next.js (Turbopack)	0.170s	1.002s	0.011s	1.017s	0.847s	10	1.24x
🌐 Redis	Next.js (Turbopack)	0.182s	1.000s	0.002s	1.007s	0.825s	10	1.33x
💻 Local	Nitro	0.212s (+7.3% 🔺)	1.003s (~)	0.011s (-2.7%)	1.017s (~)	0.804s	10	1.54x
🐘 Postgres	Express	0.215s (+2.8%)	0.996s (~)	0.002s (+15.4% 🔺)	1.013s (~)	0.797s	10	1.57x
🐘 Postgres	Nitro	0.246s (+9.4% 🔺)	0.993s (~)	0.002s (+41.7% 🔺)	1.014s (~)	0.767s	10	1.79x
🌐 MongoDB	Next.js (Turbopack)	0.474s	0.979s	0.002s	1.009s	0.534s	10	3.45x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	1.606s (+1.9%)	2.405s (-8.7% 🟢)	0.006s (-26.2% 🟢)	3.035s (-2.2%)	1.429s	10	1.00x
▲ Vercel	Nitro	1.658s (+1.2%)	2.506s (-14.3% 🟢)	0.504s (+8896.4% 🔺)	3.490s (+1.5%)	1.833s	10	1.03x
▲ Vercel	Next.js (Turbopack)	1.785s (-4.5%)	3.025s (+6.5% 🔺)	0.011s (+160.5% 🔺)	3.640s (+6.9% 🔺)	1.855s	10	1.11x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

Summary

Fastest Framework by World

Winner determined by most benchmark wins

World	🥇 Fastest Framework	Wins
💻 Local	Express	12/12
🐘 Postgres	Express	7/12
▲ Vercel	Nitro	7/12

Fastest World by Framework

Winner determined by most benchmark wins

Framework	🥇 Fastest World	Wins
Express	💻 Local	6/12
Next.js (Turbopack)	🌐 Redis	9/12
Nitro	🐘 Postgres	6/12

Column Definitions

• Workflow Time: Runtime reported by workflow (completedAt - createdAt) - primary metric
• TTFB: Time to First Byte - time from workflow start until first stream byte received (stream benchmarks only)
• Slurp: Time from first byte to complete stream consumption (stream benchmarks only)
• Wall Time: Total testbench time (trigger workflow + poll for result)
• Overhead: Testbench overhead (Wall Time - Workflow Time)
• Samples: Number of benchmark iterations run
• vs Fastest: How much slower compared to the fastest configuration for this benchmark

Worlds:

• 💻 Local: In-memory filesystem world (local development)
• 🐘 Postgres: PostgreSQL database world (local development)
• ▲ Vercel: Vercel production/preview deployment
• 🌐 Turso: Community world (local development)
• 🌐 MongoDB: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Jazz: Community world (local development)

---

📋 View full workflow run

[pranaygp]

Suggested change

	'Run already finished during setup, skipping',
	'Run already finished during setup, skipping replay',

maybe this would be clearer to the user? (assuming the codepath is only for run replays)

[pranaygp]

or "skipping redundant workflow execution"

[pranaygp]

human: LGTM! great detailed state check and ty for posting that on github PR review comment

[VaguelySerious]

Noticed error logs from 409s falling through the queue, which are confusing for users. Log example:

Queue callback error: Error [WorkflowAPIError]: Cannot update workflow run wrun_01KKVYQ2TNA4PCEH9MA0SHN7DV with status 'completed'. Operation requires status 'running' or 'pending'.
at cP (.next/server/chunks/[root-of-the-server]__fd848f42._.js:63:41)
at async gA (.next/server/chunks/[root-of-the-server]__fd848f42._.js:67:3954)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:86:8)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:60:148157)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:60:146659)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:83:2206)
at async default (.next/server/chunks/[root-of-the-server]__fd848f42._.js:67:7813)
at async u7.processMessage (.next/server/chunks/[root-of-the-server]__fd848f42._.js:62:13992)
at async u7.consume (.next/server/chunks/[root-of-the-server]__fd848f42._.js:62:15197) {
cause: undefined,
status: 409,
code: undefined,
url: 'https://vercel-workflow.com/api/v1/runs/wrun_01KKVYQ2TNA4PCEH9MA0SHN7DV'
}

so I ensures we catch and gracefully return for these cases. I had Claude run an eval on whether this blocks any real use cases. Eval below:

---

Concurrent scenarios and why skipping is safe

Scenario 1: Two handlers race on run_started

Two queue messages for the same run arrive while it's still pending. Both handlers call world.runs.get(), see pending, and try to create a run_started event.

Handler A: runs.get() → pending → events.create(run_started) → ✅ succeeds → continues replay loop
Handler B: runs.get() → pending → events.create(run_started) → ❌ 409 → returns early

Why it's safe: Handler A transitions the run to running and enters the replay loop. The replay loop either completes the workflow or suspends it with queued continuations. Handler B's early exit is a no-op — Handler A guarantees progress.

Location:runtime.ts catch block after run_started event creation.

Scenario 2: Two handlers race on run_completed

Two concurrent replay loops both reach the end of the workflow and try to create run_completed.

Handler A: events.create(run_completed) → ✅ succeeds → returns
Handler B: events.create(run_completed) → ❌ 409 → logs info → returns

Why it's safe: The run is completed. Both handlers return. No continuation is needed.

Location:runtime.ts catch block after run_completed event creation.

Scenario 3: Two handlers race on run_failed

Same as Scenario 2 but for failure. Both handlers detect an error in user code and try to fail the run.

Handler A: events.create(run_failed) → ✅ succeeds → returns
Handler B: events.create(run_failed) → ❌ 409 → logs info → returns

Why it's safe: The run is failed. Both handlers return. No continuation is needed.

Location:runtime.ts catch block after run_failed event creation.

Scenario 4: Run completes while another handler creates hooks

Handler A completes the workflow. Handler B (from an earlier queue message) is still in the suspension handler creating hook events.

Handler A: events.create(run_completed) → ✅ run is now terminal
Handler B: events.create(hook_created) → ❌ 409 (run terminal) → logs info, continues
→ returns from handleSuspension with pendingSteps
→ tries to execute inline step
→ events.create(step_started) → ❌ 410 (run gone) → returns { type: 'gone' }
→ caller returns

Why it's safe: Handler A already completed the workflow. Handler B's hook creation fails, and if any steps were queued, the step executor receives 410 on step_started and exits gracefully. The 410 on step_started is the safety net — even if the suspension handler returns pending steps after a 409, the step executor won't make progress on a finished run.

Location:suspension-handler.ts catch blocks for hook_created and hook_disposed; step-executor.ts 410 handling on step_started.

Scenario 5: Two handlers race on wait_completed

During the replay loop, both handlers see an elapsed wait and try to complete it.

Handler A: events.create(wait_completed) → ✅ succeeds → adds event to cache → continues replay
Handler B: events.create(wait_completed) → ❌ 409 → continue (skip this wait) → continues replay

Why it's safe: Both handlers continue their replay loops. The event log is the same for both (the winning handler's wait_completed is visible to future reads). Subsequent replay from either handler converges to the same state because replay is deterministic and event-sourced.

Location:runtime.ts wait completion loop with continue on 409.

Scenario 6: Two handlers race on step_completed

Two handlers execute the same step concurrently (e.g., both received the step's queue message). Both finish execution and try to create step_completed.

Handler A: events.create(step_completed) → ✅ succeeds → queues workflow continuation → returns
Handler B: events.create(step_completed) → ❌ 409 → returns WITHOUT queuing continuation

Why it's safe: Handler A queues the workflow continuation. Handler B does not — this is correct because only one continuation should be queued per step completion. If both queued, there would be redundant replay (which is safe but wasteful).

Location:step-executor.ts catch on step_completed event creation.

Scenario 7: Two handlers race on step_started

Two handlers try to start the same step. Handler A wins; Handler B gets 409 because the step transitioned to a terminal state.

Handler A: events.create(step_started) → ✅ succeeds → executes step → queues continuation
Handler B: events.create(step_started) → ❌ 409 (step terminal) → returns { type: 'skipped' }
→ caller queues workflow continuation (runtime.ts replay loop)

Why it's safe: Both handlers ensure workflow continuation — Handler A via step completion, Handler B via the replay loop re-queuing the workflow. The redundant replay is safe because event-sourced replay is convergent.

Location:step-executor.ts 409 handling on step_started.

Scenario 8: step_created/wait_created duplicate

During suspension handling, two concurrent replays both try to create the same step or wait event.

Handler A: events.create(step_created) → ✅ succeeds
Handler B: events.create(step_created) → ❌ 409 → logs info → continues

Why it's safe: The step/wait entity already exists. The suspension handler continues processing other items. The step will be executed by whichever handler picks it up from the queue.

Location:suspension-handler.ts catch blocks for step_created and wait_created.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
example-nextjs-workflow-turbopack	Ready	Preview, Comment	Mar 17, 2026 5:00pm
example-nextjs-workflow-webpack	Ready	Preview, Comment	Mar 17, 2026 5:00pm
example-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-astro-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-express-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-fastify-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-hono-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-nitro-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-nuxt-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-sveltekit-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-vite-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workflow-docs	Ready	Preview, Comment, Open in v0	Mar 17, 2026 5:00pm
workflow-nest	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workflow-swc-playground	Ready	Preview, Comment	Mar 17, 2026 5:00pm

[changeset-bot]

🦋 Changeset detected

Latest commit: 9668d09

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 16 packages

Name	Type
@workflow/core	Patch
@workflow/builders	Patch
@workflow/cli	Patch
@workflow/next	Patch
@workflow/nitro	Patch
@workflow/vitest	Patch
@workflow/web-shared	Patch
workflow	Patch
@workflow/world-testing	Patch
@workflow/astro	Patch
@workflow/nest	Patch
@workflow/rollup	Patch
@workflow/sveltekit	Patch
@workflow/vite	Patch
@workflow/nuxt	Patch
@workflow/ai	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

🧪 E2E Test Results

❌ Some tests failed

Summary

	Passed	Failed	Skipped	Total
✅ ▲ Vercel Production	747	0	67	814
✅ 💻 Local Development	770	0	118	888
✅ 📦 Local Production	770	0	118	888
✅ 🐘 Local Postgres	770	0	118	888
✅ 🪟 Windows	71	0	3	74
❌ 🌍 Community Worlds	116	55	15	186
✅ 📋 Other	195	0	27	222
Total	3439	55	466	3960

❌ Failed Tests

🌍 Community Worlds (55 failed)

mongodb (3 failed):

• hookWorkflow is not resumable via public webhook endpoint
• webhookWorkflow
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously

redis (2 failed):

• hookWorkflow is not resumable via public webhook endpoint
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously

turso (50 failed):

• addTenWorkflow
• addTenWorkflow
• wellKnownAgentWorkflow (.well-known/agent)
• should work with react rendering in step
• promiseAllWorkflow
• promiseRaceWorkflow
• promiseAnyWorkflow
• importedStepOnlyWorkflow
• hookWorkflow
• hookWorkflow is not resumable via public webhook endpoint
• webhookWorkflow
• sleepingWorkflow
• parallelSleepWorkflow
• nullByteWorkflow
• workflowAndStepMetadataWorkflow
• fetchWorkflow
• promiseRaceStressTestWorkflow
• error handling error propagation workflow errors nested function calls preserve message and stack trace
• error handling error propagation workflow errors cross-file imports preserve message and stack trace
• error handling error propagation step errors basic step error preserves message and stack trace
• error handling error propagation step errors cross-file step error preserves message and function names in stack
• error handling retry behavior regular Error retries until success
• error handling retry behavior FatalError fails immediately without retries
• error handling retry behavior RetryableError respects custom retryAfter delay
• error handling retry behavior maxRetries=0 disables retries
• error handling catchability FatalError can be caught and detected with FatalError.is()
• hookCleanupTestWorkflow - hook token reuse after workflow completion
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously
• hookDisposeTestWorkflow - hook token reuse after explicit disposal while workflow still running
• stepFunctionPassingWorkflow - step function references can be passed as arguments (without closure vars)
• stepFunctionWithClosureWorkflow - step function with closure variables passed as argument
• closureVariableWorkflow - nested step functions with closure variables
• spawnWorkflowFromStepWorkflow - spawning a child workflow using start() inside a step
• health check (queue-based) - workflow and step endpoints respond to health check messages
• pathsAliasWorkflow - TypeScript path aliases resolve correctly
• Calculator.calculate - static workflow method using static step methods from another class
• AllInOneService.processNumber - static workflow method using sibling static step methods
• ChainableService.processWithThis - static step methods using this to reference the class
• thisSerializationWorkflow - step function invoked with .call() and .apply()
• customSerializationWorkflow - custom class serialization with WORKFLOW_SERIALIZE/WORKFLOW_DESERIALIZE
• instanceMethodStepWorkflow - instance methods with "use step" directive
• crossContextSerdeWorkflow - classes defined in step code are deserializable in workflow context
• stepFunctionAsStartArgWorkflow - step function reference passed as start() argument
• cancelRun - cancelling a running workflow
• cancelRun via CLI - cancelling a running workflow
• pages router addTenWorkflow via pages router
• pages router promiseAllWorkflow via pages router
• pages router sleepingWorkflow via pages router
• hookWithSleepWorkflow - hook payloads delivered correctly with concurrent sleep
• sleepWithSequentialStepsWorkflow - sequential steps work with concurrent sleep (control)

Details by Category

✅ ▲ Vercel Production

App	Passed	Failed	Skipped
✅ astro	67	0	7
✅ example	67	0	7
✅ express	67	0	7
✅ fastify	67	0	7
✅ hono	67	0	7
✅ nextjs-turbopack	72	0	2
✅ nextjs-webpack	72	0	2
✅ nitro	67	0	7
✅ nuxt	67	0	7
✅ sveltekit	67	0	7
✅ vite	67	0	7

✅ 💻 Local Development

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 📦 Local Production

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 🐘 Local Postgres

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 🪟 Windows

App	Passed	Failed	Skipped
✅ nextjs-turbopack	71	0	3

❌ 🌍 Community Worlds

App	Passed	Failed	Skipped
✅ mongodb-dev	3	0	2
❌ mongodb	51	3	3
✅ redis-dev	3	0	2
❌ redis	52	2	3
✅ turso-dev	3	0	2
❌ turso	4	50	3

✅ 📋 Other

App	Passed	Failed	Skipped
✅ e2e-local-dev-nest-stable	65	0	9
✅ e2e-local-postgres-nest-stable	65	0	9
✅ e2e-local-prod-nest-stable	65	0	9

---

📋 View full workflow run

[github-actions]

📊 Benchmark Results

📈 Comparing against baseline from main branch. Green 🟢 = faster, Red 🔺 = slower.

workflow with no steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	0.038s (-15.7% 🟢)	1.005s (~)	0.968s	10	1.00x
💻 Local	Nitro	0.046s (+6.7% 🔺)	1.006s (~)	0.959s	10	1.23x
💻 Local	Next.js (Turbopack)	0.051s	1.006s	0.955s	10	1.36x
🌐 Redis	Next.js (Turbopack)	0.054s	1.005s	0.951s	10	1.44x
🐘 Postgres	Nitro	0.059s (-4.7%)	1.011s (~)	0.952s	10	1.56x
🐘 Postgres	Express	0.059s (-14.8% 🟢)	1.011s (~)	0.952s	10	1.57x
🌐 MongoDB	Next.js (Turbopack)	0.079s	1.008s	0.929s	10	2.11x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	0.465s (-33.5% 🟢)	2.036s (-22.7% 🟢)	1.571s	10	1.00x
▲ Vercel	Express	0.504s (-20.0% 🟢)	2.286s (-14.4% 🟢)	1.781s	10	1.08x
▲ Vercel	Next.js (Turbopack)	0.511s (-17.6% 🟢)	2.559s (+13.2% 🔺)	2.049s	10	1.10x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

workflow with 1 step

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.093s (-3.3%)	2.006s (~)	0.912s	10	1.00x
🌐 Redis	Next.js (Turbopack)	1.122s	2.006s	0.885s	10	1.03x
💻 Local	Nitro	1.130s (~)	2.006s (~)	0.877s	10	1.03x
💻 Local	Next.js (Turbopack)	1.130s	2.006s	0.876s	10	1.03x
🐘 Postgres	Express	1.153s (+0.7%)	2.014s (~)	0.860s	10	1.05x
🐘 Postgres	Nitro	1.158s (+1.1%)	2.012s (~)	0.855s	10	1.06x
🌐 MongoDB	Next.js (Turbopack)	1.307s	2.008s	0.700s	10	1.20x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.040s (-3.0%)	3.224s (-4.4%)	1.184s	10	1.00x
▲ Vercel	Express	2.053s (+0.9%)	3.735s (+12.4% 🔺)	1.682s	10	1.01x
▲ Vercel	Next.js (Turbopack)	2.168s (-2.0%)	3.924s (+13.8% 🔺)	1.756s	10	1.06x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

workflow with 10 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	10.615s (-2.8%)	11.023s (~)	0.408s	3	1.00x
🌐 Redis	Next.js (Turbopack)	10.784s	11.023s	0.239s	3	1.02x
💻 Local	Next.js (Turbopack)	10.812s	11.023s	0.211s	3	1.02x
🐘 Postgres	Nitro	10.927s (~)	11.042s (~)	0.115s	3	1.03x
🐘 Postgres	Express	10.970s (~)	11.375s (~)	0.405s	3	1.03x
💻 Local	Nitro	10.976s (+0.8%)	11.024s (~)	0.048s	3	1.03x
🌐 MongoDB	Next.js (Turbopack)	12.203s	13.015s	0.812s	3	1.15x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	16.779s (-7.4% 🟢)	18.655s (-5.6% 🟢)	1.876s	2	1.00x
▲ Vercel	Nitro	16.868s (+1.3%)	18.197s (-1.3%)	1.329s	2	1.01x
▲ Vercel	Next.js (Turbopack)	17.826s (+0.9%)	19.468s (+1.2%)	1.641s	2	1.06x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

workflow with 25 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	26.743s (-3.1%)	27.051s (-3.6%)	0.308s	3	1.00x
🌐 Redis	Next.js (Turbopack)	26.760s	27.049s	0.289s	3	1.00x
💻 Local	Next.js (Turbopack)	27.128s	28.053s	0.925s	3	1.01x
🐘 Postgres	Express	27.240s (~)	28.066s (~)	0.826s	3	1.02x
🐘 Postgres	Nitro	27.277s (~)	28.065s (~)	0.788s	3	1.02x
💻 Local	Nitro	27.583s (~)	28.053s (~)	0.470s	3	1.03x
🌐 MongoDB	Next.js (Turbopack)	30.482s	31.045s	0.564s	2	1.14x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	46.028s (+2.9%)	47.737s (+4.3%)	1.709s	2	1.00x
▲ Vercel	Next.js (Turbopack)	50.377s (+6.6% 🔺)	52.253s (+8.0% 🔺)	1.877s	2	1.09x
▲ Vercel	Express	53.606s (+21.0% 🔺)	55.319s (+22.1% 🔺)	1.713s	2	1.16x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

workflow with 50 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	53.541s	54.093s	0.552s	2	1.00x
🐘 Postgres	Express	54.300s (~)	55.094s (~)	0.793s	2	1.01x
🐘 Postgres	Nitro	54.476s (~)	55.105s (~)	0.629s	2	1.02x
💻 Local	Express	54.851s (-3.3%)	55.101s (-3.5%)	0.250s	2	1.02x
💻 Local	Next.js (Turbopack)	55.951s	56.100s	0.149s	2	1.05x
💻 Local	Nitro	56.657s (~)	57.106s (~)	0.450s	2	1.06x
🌐 MongoDB	Next.js (Turbopack)	60.675s	61.065s	0.390s	2	1.13x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	94.771s (-4.4%)	96.975s (-4.2%)	2.204s	1	1.00x
▲ Vercel	Nitro	96.605s (~)	98.384s (~)	1.779s	1	1.02x
▲ Vercel	Next.js (Turbopack)	98.753s (-3.9%)	100.796s (-2.8%)	2.043s	1	1.04x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

Promise.all with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	1.345s	2.006s	0.662s	15	1.00x
🐘 Postgres	Express	1.366s (-4.0%)	2.011s (~)	0.645s	15	1.02x
💻 Local	Express	1.467s (-5.2% 🟢)	2.006s (~)	0.538s	15	1.09x
💻 Local	Nitro	1.498s (-1.1%)	2.005s (~)	0.507s	15	1.11x
💻 Local	Next.js (Turbopack)	1.547s	2.006s	0.459s	15	1.15x
🌐 MongoDB	Next.js (Turbopack)	2.146s	3.009s	0.863s	10	1.60x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-
🐘 Postgres	Nitro	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.312s (-4.2%)	3.407s (-4.8%)	1.095s	9	1.00x
▲ Vercel	Next.js (Turbopack)	2.641s (-1.3%)	4.038s (+7.5% 🔺)	1.397s	8	1.14x
▲ Vercel	Express	2.648s (+5.0%)	4.138s (+13.0% 🔺)	1.490s	8	1.15x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Promise.all with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	2.567s	3.008s	0.440s	10	1.00x
🐘 Postgres	Nitro	2.595s (~)	3.014s (~)	0.419s	10	1.01x
💻 Local	Express	2.600s (-14.1% 🟢)	3.007s (-15.6% 🟢)	0.407s	10	1.01x
🐘 Postgres	Express	2.615s (~)	3.015s (~)	0.399s	10	1.02x
💻 Local	Next.js (Turbopack)	3.041s	3.760s	0.718s	8	1.18x
💻 Local	Nitro	3.121s (+7.5% 🔺)	3.565s (+11.1% 🔺)	0.444s	9	1.22x
🌐 MongoDB	Next.js (Turbopack)	4.747s	5.179s	0.432s	6	1.85x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Next.js (Turbopack)	2.799s (+7.7% 🔺)	4.348s (+21.1% 🔺)	1.550s	8	1.00x
▲ Vercel	Nitro	3.520s (+30.2% 🔺)	4.708s (+23.4% 🔺)	1.188s	7	1.26x
▲ Vercel	Express	3.663s (+47.9% 🔺)	5.183s (+40.8% 🔺)	1.520s	8	1.31x

🔍 Observability: Next.js (Turbopack) | Nitro | Express

Promise.all with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	4.027s (+1.4%)	4.588s (+3.2%)	0.561s	7	1.00x
🐘 Postgres	Nitro	4.032s (+0.7%)	4.590s (+3.1%)	0.558s	7	1.00x
🌐 Redis	Next.js (Turbopack)	4.081s	5.012s	0.931s	6	1.01x
💻 Local	Express	6.787s (-15.7% 🟢)	7.014s (-20.1% 🟢)	0.227s	5	1.69x
💻 Local	Next.js (Turbopack)	7.888s	8.517s	0.629s	4	1.96x
💻 Local	Nitro	8.207s (-1.6%)	8.521s (-5.5% 🟢)	0.314s	4	2.04x
🌐 MongoDB	Next.js (Turbopack)	10.067s	10.684s	0.617s	3	2.50x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Next.js (Turbopack)	5.520s (+44.3% 🔺)	7.146s (+37.1% 🔺)	1.626s	5	1.00x
▲ Vercel	Nitro	7.009s (+153.3% 🔺)	8.352s (+119.6% 🔺)	1.343s	4	1.27x
▲ Vercel	Express	10.371s (+264.0% 🔺)	12.346s (+209.2% 🔺)	1.975s	3	1.88x

🔍 Observability: Next.js (Turbopack) | Nitro | Express

Promise.race with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	1.303s	2.006s	0.703s	15	1.00x
🐘 Postgres	Nitro	1.422s (-1.3%)	2.011s (~)	0.589s	15	1.09x
🐘 Postgres	Express	1.442s (-1.5%)	2.011s (-3.2%)	0.570s	15	1.11x
💻 Local	Express	1.477s (-2.9%)	2.005s (~)	0.528s	15	1.13x
💻 Local	Nitro	1.523s (-1.1%)	2.006s (~)	0.483s	15	1.17x
💻 Local	Next.js (Turbopack)	1.564s	2.073s	0.509s	15	1.20x
🌐 MongoDB	Next.js (Turbopack)	2.174s	3.009s	0.835s	10	1.67x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.131s (-0.9%)	3.290s (-1.5%)	1.159s	10	1.00x
▲ Vercel	Next.js (Turbopack)	2.268s (-27.5% 🟢)	3.823s (-24.1% 🟢)	1.555s	8	1.06x
▲ Vercel	Express	2.591s (+17.2% 🔺)	4.029s (+17.2% 🔺)	1.438s	8	1.22x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Promise.race with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	2.526s (-3.2%)	3.012s (~)	0.486s	10	1.00x
🌐 Redis	Next.js (Turbopack)	2.532s	3.008s	0.477s	10	1.00x
🐘 Postgres	Nitro	2.730s (+7.4% 🔺)	3.015s (~)	0.285s	10	1.08x
💻 Local	Express	2.797s (-9.8% 🟢)	3.108s (-17.3% 🟢)	0.311s	10	1.11x
💻 Local	Next.js (Turbopack)	3.030s	3.760s	0.729s	8	1.20x
💻 Local	Nitro	3.124s (-1.8%)	3.886s (~)	0.762s	8	1.24x
🌐 MongoDB	Next.js (Turbopack)	4.728s	5.177s	0.449s	6	1.87x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.270s (-29.0% 🟢)	3.269s (-37.5% 🟢)	0.998s	10	1.00x
▲ Vercel	Express	2.378s (-32.1% 🟢)	3.827s (-17.2% 🟢)	1.449s	8	1.05x
▲ Vercel	Next.js (Turbopack)	3.093s (-5.5% 🟢)	4.610s (~)	1.517s	7	1.36x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

Promise.race with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	3.975s (-3.5%)	4.590s (~)	0.615s	7	1.00x
🐘 Postgres	Express	4.039s (+2.3%)	4.590s (+3.2%)	0.551s	7	1.02x
🌐 Redis	Next.js (Turbopack)	4.198s	4.725s	0.527s	7	1.06x
💻 Local	Express	7.733s (-14.1% 🟢)	8.267s (-10.8% 🟢)	0.533s	4	1.95x
💻 Local	Next.js (Turbopack)	8.255s	8.766s	0.512s	4	2.08x
💻 Local	Nitro	8.519s (+0.8%)	9.275s (+2.8%)	0.756s	4	2.14x
🌐 MongoDB	Next.js (Turbopack)	10.032s	10.680s	0.648s	3	2.52x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.724s (-27.3% 🟢)	3.647s (-28.7% 🟢)	0.923s	9	1.00x
▲ Vercel	Next.js (Turbopack)	3.571s (-3.3%)	5.369s (+9.7% 🔺)	1.798s	6	1.31x
▲ Vercel	Express	3.789s (+1.4%)	5.100s (-0.7%)	1.311s	6	1.39x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Stream Benchmarks(includes TTFB metrics)

workflow with stream

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	0.138s (-32.2% 🟢)	1.003s (~)	0.009s (-21.6% 🟢)	1.015s (~)	0.877s	10	1.00x
💻 Local	Next.js (Turbopack)	0.170s	1.002s	0.011s	1.017s	0.847s	10	1.24x
🌐 Redis	Next.js (Turbopack)	0.182s	1.000s	0.002s	1.007s	0.825s	10	1.33x
💻 Local	Nitro	0.212s (+7.3% 🔺)	1.003s (~)	0.011s (-2.7%)	1.017s (~)	0.804s	10	1.54x
🐘 Postgres	Express	0.215s (+2.8%)	0.996s (~)	0.002s (+15.4% 🔺)	1.013s (~)	0.797s	10	1.57x
🐘 Postgres	Nitro	0.246s (+9.4% 🔺)	0.993s (~)	0.002s (+41.7% 🔺)	1.014s (~)	0.767s	10	1.79x
🌐 MongoDB	Next.js (Turbopack)	0.474s	0.979s	0.002s	1.009s	0.534s	10	3.45x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	1.606s (+1.9%)	2.405s (-8.7% 🟢)	0.006s (-26.2% 🟢)	3.035s (-2.2%)	1.429s	10	1.00x
▲ Vercel	Nitro	1.658s (+1.2%)	2.506s (-14.3% 🟢)	0.504s (+8896.4% 🔺)	3.490s (+1.5%)	1.833s	10	1.03x
▲ Vercel	Next.js (Turbopack)	1.785s (-4.5%)	3.025s (+6.5% 🔺)	0.011s (+160.5% 🔺)	3.640s (+6.9% 🔺)	1.855s	10	1.11x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

Summary

Fastest Framework by World

Winner determined by most benchmark wins

World	🥇 Fastest Framework	Wins
💻 Local	Express	12/12
🐘 Postgres	Express	7/12
▲ Vercel	Nitro	7/12

Fastest World by Framework

Winner determined by most benchmark wins

Framework	🥇 Fastest World	Wins
Express	💻 Local	6/12
Next.js (Turbopack)	🌐 Redis	9/12
Nitro	🐘 Postgres	6/12

Column Definitions

• Workflow Time: Runtime reported by workflow (completedAt - createdAt) - primary metric
• TTFB: Time to First Byte - time from workflow start until first stream byte received (stream benchmarks only)
• Slurp: Time from first byte to complete stream consumption (stream benchmarks only)
• Wall Time: Total testbench time (trigger workflow + poll for result)
• Overhead: Testbench overhead (Wall Time - Workflow Time)
• Samples: Number of benchmark iterations run
• vs Fastest: How much slower compared to the fastest configuration for this benchmark

Worlds:

• 💻 Local: In-memory filesystem world (local development)
• 🐘 Postgres: PostgreSQL database world (local development)
• ▲ Vercel: Vercel production/preview deployment
• 🌐 Turso: Community world (local development)
• 🌐 MongoDB: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Jazz: Community world (local development)

---

📋 View full workflow run

[pranaygp]

Suggested change

	'Run already finished during setup, skipping',
	'Run already finished during setup, skipping replay',

maybe this would be clearer to the user? (assuming the codepath is only for run replays)

[pranaygp]

or "skipping redundant workflow execution"

[pranaygp]

human: LGTM! great detailed state check and ty for posting that on github PR review comment

[VaguelySerious]

Noticed error logs from 409s falling through the queue, which are confusing for users. Log example:

Queue callback error: Error [WorkflowAPIError]: Cannot update workflow run wrun_01KKVYQ2TNA4PCEH9MA0SHN7DV with status 'completed'. Operation requires status 'running' or 'pending'.
at cP (.next/server/chunks/[root-of-the-server]__fd848f42._.js:63:41)
at async gA (.next/server/chunks/[root-of-the-server]__fd848f42._.js:67:3954)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:86:8)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:60:148157)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:60:146659)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:83:2206)
at async default (.next/server/chunks/[root-of-the-server]__fd848f42._.js:67:7813)
at async u7.processMessage (.next/server/chunks/[root-of-the-server]__fd848f42._.js:62:13992)
at async u7.consume (.next/server/chunks/[root-of-the-server]__fd848f42._.js:62:15197) {
cause: undefined,
status: 409,
code: undefined,
url: 'https://vercel-workflow.com/api/v1/runs/wrun_01KKVYQ2TNA4PCEH9MA0SHN7DV'
}

so I ensures we catch and gracefully return for these cases. I had Claude run an eval on whether this blocks any real use cases. Eval below:

---

Concurrent scenarios and why skipping is safe

Scenario 1: Two handlers race on run_started

Two queue messages for the same run arrive while it's still pending. Both handlers call world.runs.get(), see pending, and try to create a run_started event.

Handler A: runs.get() → pending → events.create(run_started) → ✅ succeeds → continues replay loop
Handler B: runs.get() → pending → events.create(run_started) → ❌ 409 → returns early

Why it's safe: Handler A transitions the run to running and enters the replay loop. The replay loop either completes the workflow or suspends it with queued continuations. Handler B's early exit is a no-op — Handler A guarantees progress.

Location:runtime.ts catch block after run_started event creation.

Scenario 2: Two handlers race on run_completed

Two concurrent replay loops both reach the end of the workflow and try to create run_completed.

Handler A: events.create(run_completed) → ✅ succeeds → returns
Handler B: events.create(run_completed) → ❌ 409 → logs info → returns

Why it's safe: The run is completed. Both handlers return. No continuation is needed.

Location:runtime.ts catch block after run_completed event creation.

Scenario 3: Two handlers race on run_failed

Same as Scenario 2 but for failure. Both handlers detect an error in user code and try to fail the run.

Handler A: events.create(run_failed) → ✅ succeeds → returns
Handler B: events.create(run_failed) → ❌ 409 → logs info → returns

Why it's safe: The run is failed. Both handlers return. No continuation is needed.

Location:runtime.ts catch block after run_failed event creation.

Scenario 4: Run completes while another handler creates hooks

Handler A completes the workflow. Handler B (from an earlier queue message) is still in the suspension handler creating hook events.

Handler A: events.create(run_completed) → ✅ run is now terminal
Handler B: events.create(hook_created) → ❌ 409 (run terminal) → logs info, continues
→ returns from handleSuspension with pendingSteps
→ tries to execute inline step
→ events.create(step_started) → ❌ 410 (run gone) → returns { type: 'gone' }
→ caller returns

Why it's safe: Handler A already completed the workflow. Handler B's hook creation fails, and if any steps were queued, the step executor receives 410 on step_started and exits gracefully. The 410 on step_started is the safety net — even if the suspension handler returns pending steps after a 409, the step executor won't make progress on a finished run.

Location:suspension-handler.ts catch blocks for hook_created and hook_disposed; step-executor.ts 410 handling on step_started.

Scenario 5: Two handlers race on wait_completed

During the replay loop, both handlers see an elapsed wait and try to complete it.

Handler A: events.create(wait_completed) → ✅ succeeds → adds event to cache → continues replay
Handler B: events.create(wait_completed) → ❌ 409 → continue (skip this wait) → continues replay

Why it's safe: Both handlers continue their replay loops. The event log is the same for both (the winning handler's wait_completed is visible to future reads). Subsequent replay from either handler converges to the same state because replay is deterministic and event-sourced.

Location:runtime.ts wait completion loop with continue on 409.

Scenario 6: Two handlers race on step_completed

Two handlers execute the same step concurrently (e.g., both received the step's queue message). Both finish execution and try to create step_completed.

Handler A: events.create(step_completed) → ✅ succeeds → queues workflow continuation → returns
Handler B: events.create(step_completed) → ❌ 409 → returns WITHOUT queuing continuation

Why it's safe: Handler A queues the workflow continuation. Handler B does not — this is correct because only one continuation should be queued per step completion. If both queued, there would be redundant replay (which is safe but wasteful).

Location:step-executor.ts catch on step_completed event creation.

Scenario 7: Two handlers race on step_started

Two handlers try to start the same step. Handler A wins; Handler B gets 409 because the step transitioned to a terminal state.

Handler A: events.create(step_started) → ✅ succeeds → executes step → queues continuation
Handler B: events.create(step_started) → ❌ 409 (step terminal) → returns { type: 'skipped' }
→ caller queues workflow continuation (runtime.ts replay loop)

Why it's safe: Both handlers ensure workflow continuation — Handler A via step completion, Handler B via the replay loop re-queuing the workflow. The redundant replay is safe because event-sourced replay is convergent.

Location:step-executor.ts 409 handling on step_started.

Scenario 8: step_created/wait_created duplicate

During suspension handling, two concurrent replays both try to create the same step or wait event.

Handler A: events.create(step_created) → ✅ succeeds
Handler B: events.create(step_created) → ❌ 409 → logs info → continues

Why it's safe: The step/wait entity already exists. The suspension handler continues processing other items. The step will be executed by whichever handler picks it up from the queue.

Location:suspension-handler.ts catch blocks for step_created and wait_created.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
example-nextjs-workflow-turbopack	Ready	Preview, Comment	Mar 17, 2026 5:00pm
example-nextjs-workflow-webpack	Ready	Preview, Comment	Mar 17, 2026 5:00pm
example-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-astro-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-express-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-fastify-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-hono-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-nitro-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-nuxt-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-sveltekit-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-vite-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workflow-docs	Ready	Preview, Comment, Open in v0	Mar 17, 2026 5:00pm
workflow-nest	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workflow-swc-playground	Ready	Preview, Comment	Mar 17, 2026 5:00pm

[changeset-bot]

🦋 Changeset detected

Latest commit: 9668d09

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 16 packages

Name	Type
@workflow/core	Patch
@workflow/builders	Patch
@workflow/cli	Patch
@workflow/next	Patch
@workflow/nitro	Patch
@workflow/vitest	Patch
@workflow/web-shared	Patch
workflow	Patch
@workflow/world-testing	Patch
@workflow/astro	Patch
@workflow/nest	Patch
@workflow/rollup	Patch
@workflow/sveltekit	Patch
@workflow/vite	Patch
@workflow/nuxt	Patch
@workflow/ai	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

🧪 E2E Test Results

❌ Some tests failed

Summary

	Passed	Failed	Skipped	Total
✅ ▲ Vercel Production	747	0	67	814
✅ 💻 Local Development	770	0	118	888
✅ 📦 Local Production	770	0	118	888
✅ 🐘 Local Postgres	770	0	118	888
✅ 🪟 Windows	71	0	3	74
❌ 🌍 Community Worlds	116	55	15	186
✅ 📋 Other	195	0	27	222
Total	3439	55	466	3960

❌ Failed Tests

🌍 Community Worlds (55 failed)

mongodb (3 failed):

• hookWorkflow is not resumable via public webhook endpoint
• webhookWorkflow
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously

redis (2 failed):

• hookWorkflow is not resumable via public webhook endpoint
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously

turso (50 failed):

• addTenWorkflow
• addTenWorkflow
• wellKnownAgentWorkflow (.well-known/agent)
• should work with react rendering in step
• promiseAllWorkflow
• promiseRaceWorkflow
• promiseAnyWorkflow
• importedStepOnlyWorkflow
• hookWorkflow
• hookWorkflow is not resumable via public webhook endpoint
• webhookWorkflow
• sleepingWorkflow
• parallelSleepWorkflow
• nullByteWorkflow
• workflowAndStepMetadataWorkflow
• fetchWorkflow
• promiseRaceStressTestWorkflow
• error handling error propagation workflow errors nested function calls preserve message and stack trace
• error handling error propagation workflow errors cross-file imports preserve message and stack trace
• error handling error propagation step errors basic step error preserves message and stack trace
• error handling error propagation step errors cross-file step error preserves message and function names in stack
• error handling retry behavior regular Error retries until success
• error handling retry behavior FatalError fails immediately without retries
• error handling retry behavior RetryableError respects custom retryAfter delay
• error handling retry behavior maxRetries=0 disables retries
• error handling catchability FatalError can be caught and detected with FatalError.is()
• hookCleanupTestWorkflow - hook token reuse after workflow completion
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously
• hookDisposeTestWorkflow - hook token reuse after explicit disposal while workflow still running
• stepFunctionPassingWorkflow - step function references can be passed as arguments (without closure vars)
• stepFunctionWithClosureWorkflow - step function with closure variables passed as argument
• closureVariableWorkflow - nested step functions with closure variables
• spawnWorkflowFromStepWorkflow - spawning a child workflow using start() inside a step
• health check (queue-based) - workflow and step endpoints respond to health check messages
• pathsAliasWorkflow - TypeScript path aliases resolve correctly
• Calculator.calculate - static workflow method using static step methods from another class
• AllInOneService.processNumber - static workflow method using sibling static step methods
• ChainableService.processWithThis - static step methods using this to reference the class
• thisSerializationWorkflow - step function invoked with .call() and .apply()
• customSerializationWorkflow - custom class serialization with WORKFLOW_SERIALIZE/WORKFLOW_DESERIALIZE
• instanceMethodStepWorkflow - instance methods with "use step" directive
• crossContextSerdeWorkflow - classes defined in step code are deserializable in workflow context
• stepFunctionAsStartArgWorkflow - step function reference passed as start() argument
• cancelRun - cancelling a running workflow
• cancelRun via CLI - cancelling a running workflow
• pages router addTenWorkflow via pages router
• pages router promiseAllWorkflow via pages router
• pages router sleepingWorkflow via pages router
• hookWithSleepWorkflow - hook payloads delivered correctly with concurrent sleep
• sleepWithSequentialStepsWorkflow - sequential steps work with concurrent sleep (control)

Details by Category

✅ ▲ Vercel Production

App	Passed	Failed	Skipped
✅ astro	67	0	7
✅ example	67	0	7
✅ express	67	0	7
✅ fastify	67	0	7
✅ hono	67	0	7
✅ nextjs-turbopack	72	0	2
✅ nextjs-webpack	72	0	2
✅ nitro	67	0	7
✅ nuxt	67	0	7
✅ sveltekit	67	0	7
✅ vite	67	0	7

✅ 💻 Local Development

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 📦 Local Production

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 🐘 Local Postgres

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 🪟 Windows

App	Passed	Failed	Skipped
✅ nextjs-turbopack	71	0	3

❌ 🌍 Community Worlds

App	Passed	Failed	Skipped
✅ mongodb-dev	3	0	2
❌ mongodb	51	3	3
✅ redis-dev	3	0	2
❌ redis	52	2	3
✅ turso-dev	3	0	2
❌ turso	4	50	3

✅ 📋 Other

App	Passed	Failed	Skipped
✅ e2e-local-dev-nest-stable	65	0	9
✅ e2e-local-postgres-nest-stable	65	0	9
✅ e2e-local-prod-nest-stable	65	0	9

---

📋 View full workflow run

[github-actions]

📊 Benchmark Results

📈 Comparing against baseline from main branch. Green 🟢 = faster, Red 🔺 = slower.

workflow with no steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	0.038s (-15.7% 🟢)	1.005s (~)	0.968s	10	1.00x
💻 Local	Nitro	0.046s (+6.7% 🔺)	1.006s (~)	0.959s	10	1.23x
💻 Local	Next.js (Turbopack)	0.051s	1.006s	0.955s	10	1.36x
🌐 Redis	Next.js (Turbopack)	0.054s	1.005s	0.951s	10	1.44x
🐘 Postgres	Nitro	0.059s (-4.7%)	1.011s (~)	0.952s	10	1.56x
🐘 Postgres	Express	0.059s (-14.8% 🟢)	1.011s (~)	0.952s	10	1.57x
🌐 MongoDB	Next.js (Turbopack)	0.079s	1.008s	0.929s	10	2.11x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	0.465s (-33.5% 🟢)	2.036s (-22.7% 🟢)	1.571s	10	1.00x
▲ Vercel	Express	0.504s (-20.0% 🟢)	2.286s (-14.4% 🟢)	1.781s	10	1.08x
▲ Vercel	Next.js (Turbopack)	0.511s (-17.6% 🟢)	2.559s (+13.2% 🔺)	2.049s	10	1.10x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

workflow with 1 step

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.093s (-3.3%)	2.006s (~)	0.912s	10	1.00x
🌐 Redis	Next.js (Turbopack)	1.122s	2.006s	0.885s	10	1.03x
💻 Local	Nitro	1.130s (~)	2.006s (~)	0.877s	10	1.03x
💻 Local	Next.js (Turbopack)	1.130s	2.006s	0.876s	10	1.03x
🐘 Postgres	Express	1.153s (+0.7%)	2.014s (~)	0.860s	10	1.05x
🐘 Postgres	Nitro	1.158s (+1.1%)	2.012s (~)	0.855s	10	1.06x
🌐 MongoDB	Next.js (Turbopack)	1.307s	2.008s	0.700s	10	1.20x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.040s (-3.0%)	3.224s (-4.4%)	1.184s	10	1.00x
▲ Vercel	Express	2.053s (+0.9%)	3.735s (+12.4% 🔺)	1.682s	10	1.01x
▲ Vercel	Next.js (Turbopack)	2.168s (-2.0%)	3.924s (+13.8% 🔺)	1.756s	10	1.06x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

workflow with 10 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	10.615s (-2.8%)	11.023s (~)	0.408s	3	1.00x
🌐 Redis	Next.js (Turbopack)	10.784s	11.023s	0.239s	3	1.02x
💻 Local	Next.js (Turbopack)	10.812s	11.023s	0.211s	3	1.02x
🐘 Postgres	Nitro	10.927s (~)	11.042s (~)	0.115s	3	1.03x
🐘 Postgres	Express	10.970s (~)	11.375s (~)	0.405s	3	1.03x
💻 Local	Nitro	10.976s (+0.8%)	11.024s (~)	0.048s	3	1.03x
🌐 MongoDB	Next.js (Turbopack)	12.203s	13.015s	0.812s	3	1.15x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	16.779s (-7.4% 🟢)	18.655s (-5.6% 🟢)	1.876s	2	1.00x
▲ Vercel	Nitro	16.868s (+1.3%)	18.197s (-1.3%)	1.329s	2	1.01x
▲ Vercel	Next.js (Turbopack)	17.826s (+0.9%)	19.468s (+1.2%)	1.641s	2	1.06x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

workflow with 25 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	26.743s (-3.1%)	27.051s (-3.6%)	0.308s	3	1.00x
🌐 Redis	Next.js (Turbopack)	26.760s	27.049s	0.289s	3	1.00x
💻 Local	Next.js (Turbopack)	27.128s	28.053s	0.925s	3	1.01x
🐘 Postgres	Express	27.240s (~)	28.066s (~)	0.826s	3	1.02x
🐘 Postgres	Nitro	27.277s (~)	28.065s (~)	0.788s	3	1.02x
💻 Local	Nitro	27.583s (~)	28.053s (~)	0.470s	3	1.03x
🌐 MongoDB	Next.js (Turbopack)	30.482s	31.045s	0.564s	2	1.14x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	46.028s (+2.9%)	47.737s (+4.3%)	1.709s	2	1.00x
▲ Vercel	Next.js (Turbopack)	50.377s (+6.6% 🔺)	52.253s (+8.0% 🔺)	1.877s	2	1.09x
▲ Vercel	Express	53.606s (+21.0% 🔺)	55.319s (+22.1% 🔺)	1.713s	2	1.16x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

workflow with 50 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	53.541s	54.093s	0.552s	2	1.00x
🐘 Postgres	Express	54.300s (~)	55.094s (~)	0.793s	2	1.01x
🐘 Postgres	Nitro	54.476s (~)	55.105s (~)	0.629s	2	1.02x
💻 Local	Express	54.851s (-3.3%)	55.101s (-3.5%)	0.250s	2	1.02x
💻 Local	Next.js (Turbopack)	55.951s	56.100s	0.149s	2	1.05x
💻 Local	Nitro	56.657s (~)	57.106s (~)	0.450s	2	1.06x
🌐 MongoDB	Next.js (Turbopack)	60.675s	61.065s	0.390s	2	1.13x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	94.771s (-4.4%)	96.975s (-4.2%)	2.204s	1	1.00x
▲ Vercel	Nitro	96.605s (~)	98.384s (~)	1.779s	1	1.02x
▲ Vercel	Next.js (Turbopack)	98.753s (-3.9%)	100.796s (-2.8%)	2.043s	1	1.04x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

Promise.all with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	1.345s	2.006s	0.662s	15	1.00x
🐘 Postgres	Express	1.366s (-4.0%)	2.011s (~)	0.645s	15	1.02x
💻 Local	Express	1.467s (-5.2% 🟢)	2.006s (~)	0.538s	15	1.09x
💻 Local	Nitro	1.498s (-1.1%)	2.005s (~)	0.507s	15	1.11x
💻 Local	Next.js (Turbopack)	1.547s	2.006s	0.459s	15	1.15x
🌐 MongoDB	Next.js (Turbopack)	2.146s	3.009s	0.863s	10	1.60x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-
🐘 Postgres	Nitro	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.312s (-4.2%)	3.407s (-4.8%)	1.095s	9	1.00x
▲ Vercel	Next.js (Turbopack)	2.641s (-1.3%)	4.038s (+7.5% 🔺)	1.397s	8	1.14x
▲ Vercel	Express	2.648s (+5.0%)	4.138s (+13.0% 🔺)	1.490s	8	1.15x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Promise.all with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	2.567s	3.008s	0.440s	10	1.00x
🐘 Postgres	Nitro	2.595s (~)	3.014s (~)	0.419s	10	1.01x
💻 Local	Express	2.600s (-14.1% 🟢)	3.007s (-15.6% 🟢)	0.407s	10	1.01x
🐘 Postgres	Express	2.615s (~)	3.015s (~)	0.399s	10	1.02x
💻 Local	Next.js (Turbopack)	3.041s	3.760s	0.718s	8	1.18x
💻 Local	Nitro	3.121s (+7.5% 🔺)	3.565s (+11.1% 🔺)	0.444s	9	1.22x
🌐 MongoDB	Next.js (Turbopack)	4.747s	5.179s	0.432s	6	1.85x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Next.js (Turbopack)	2.799s (+7.7% 🔺)	4.348s (+21.1% 🔺)	1.550s	8	1.00x
▲ Vercel	Nitro	3.520s (+30.2% 🔺)	4.708s (+23.4% 🔺)	1.188s	7	1.26x
▲ Vercel	Express	3.663s (+47.9% 🔺)	5.183s (+40.8% 🔺)	1.520s	8	1.31x

🔍 Observability: Next.js (Turbopack) | Nitro | Express

Promise.all with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	4.027s (+1.4%)	4.588s (+3.2%)	0.561s	7	1.00x
🐘 Postgres	Nitro	4.032s (+0.7%)	4.590s (+3.1%)	0.558s	7	1.00x
🌐 Redis	Next.js (Turbopack)	4.081s	5.012s	0.931s	6	1.01x
💻 Local	Express	6.787s (-15.7% 🟢)	7.014s (-20.1% 🟢)	0.227s	5	1.69x
💻 Local	Next.js (Turbopack)	7.888s	8.517s	0.629s	4	1.96x
💻 Local	Nitro	8.207s (-1.6%)	8.521s (-5.5% 🟢)	0.314s	4	2.04x
🌐 MongoDB	Next.js (Turbopack)	10.067s	10.684s	0.617s	3	2.50x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Next.js (Turbopack)	5.520s (+44.3% 🔺)	7.146s (+37.1% 🔺)	1.626s	5	1.00x
▲ Vercel	Nitro	7.009s (+153.3% 🔺)	8.352s (+119.6% 🔺)	1.343s	4	1.27x
▲ Vercel	Express	10.371s (+264.0% 🔺)	12.346s (+209.2% 🔺)	1.975s	3	1.88x

🔍 Observability: Next.js (Turbopack) | Nitro | Express

Promise.race with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	1.303s	2.006s	0.703s	15	1.00x
🐘 Postgres	Nitro	1.422s (-1.3%)	2.011s (~)	0.589s	15	1.09x
🐘 Postgres	Express	1.442s (-1.5%)	2.011s (-3.2%)	0.570s	15	1.11x
💻 Local	Express	1.477s (-2.9%)	2.005s (~)	0.528s	15	1.13x
💻 Local	Nitro	1.523s (-1.1%)	2.006s (~)	0.483s	15	1.17x
💻 Local	Next.js (Turbopack)	1.564s	2.073s	0.509s	15	1.20x
🌐 MongoDB	Next.js (Turbopack)	2.174s	3.009s	0.835s	10	1.67x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.131s (-0.9%)	3.290s (-1.5%)	1.159s	10	1.00x
▲ Vercel	Next.js (Turbopack)	2.268s (-27.5% 🟢)	3.823s (-24.1% 🟢)	1.555s	8	1.06x
▲ Vercel	Express	2.591s (+17.2% 🔺)	4.029s (+17.2% 🔺)	1.438s	8	1.22x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Promise.race with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	2.526s (-3.2%)	3.012s (~)	0.486s	10	1.00x
🌐 Redis	Next.js (Turbopack)	2.532s	3.008s	0.477s	10	1.00x
🐘 Postgres	Nitro	2.730s (+7.4% 🔺)	3.015s (~)	0.285s	10	1.08x
💻 Local	Express	2.797s (-9.8% 🟢)	3.108s (-17.3% 🟢)	0.311s	10	1.11x
💻 Local	Next.js (Turbopack)	3.030s	3.760s	0.729s	8	1.20x
💻 Local	Nitro	3.124s (-1.8%)	3.886s (~)	0.762s	8	1.24x
🌐 MongoDB	Next.js (Turbopack)	4.728s	5.177s	0.449s	6	1.87x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.270s (-29.0% 🟢)	3.269s (-37.5% 🟢)	0.998s	10	1.00x
▲ Vercel	Express	2.378s (-32.1% 🟢)	3.827s (-17.2% 🟢)	1.449s	8	1.05x
▲ Vercel	Next.js (Turbopack)	3.093s (-5.5% 🟢)	4.610s (~)	1.517s	7	1.36x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

Promise.race with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	3.975s (-3.5%)	4.590s (~)	0.615s	7	1.00x
🐘 Postgres	Express	4.039s (+2.3%)	4.590s (+3.2%)	0.551s	7	1.02x
🌐 Redis	Next.js (Turbopack)	4.198s	4.725s	0.527s	7	1.06x
💻 Local	Express	7.733s (-14.1% 🟢)	8.267s (-10.8% 🟢)	0.533s	4	1.95x
💻 Local	Next.js (Turbopack)	8.255s	8.766s	0.512s	4	2.08x
💻 Local	Nitro	8.519s (+0.8%)	9.275s (+2.8%)	0.756s	4	2.14x
🌐 MongoDB	Next.js (Turbopack)	10.032s	10.680s	0.648s	3	2.52x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.724s (-27.3% 🟢)	3.647s (-28.7% 🟢)	0.923s	9	1.00x
▲ Vercel	Next.js (Turbopack)	3.571s (-3.3%)	5.369s (+9.7% 🔺)	1.798s	6	1.31x
▲ Vercel	Express	3.789s (+1.4%)	5.100s (-0.7%)	1.311s	6	1.39x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Stream Benchmarks(includes TTFB metrics)

workflow with stream

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	0.138s (-32.2% 🟢)	1.003s (~)	0.009s (-21.6% 🟢)	1.015s (~)	0.877s	10	1.00x
💻 Local	Next.js (Turbopack)	0.170s	1.002s	0.011s	1.017s	0.847s	10	1.24x
🌐 Redis	Next.js (Turbopack)	0.182s	1.000s	0.002s	1.007s	0.825s	10	1.33x
💻 Local	Nitro	0.212s (+7.3% 🔺)	1.003s (~)	0.011s (-2.7%)	1.017s (~)	0.804s	10	1.54x
🐘 Postgres	Express	0.215s (+2.8%)	0.996s (~)	0.002s (+15.4% 🔺)	1.013s (~)	0.797s	10	1.57x
🐘 Postgres	Nitro	0.246s (+9.4% 🔺)	0.993s (~)	0.002s (+41.7% 🔺)	1.014s (~)	0.767s	10	1.79x
🌐 MongoDB	Next.js (Turbopack)	0.474s	0.979s	0.002s	1.009s	0.534s	10	3.45x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	1.606s (+1.9%)	2.405s (-8.7% 🟢)	0.006s (-26.2% 🟢)	3.035s (-2.2%)	1.429s	10	1.00x
▲ Vercel	Nitro	1.658s (+1.2%)	2.506s (-14.3% 🟢)	0.504s (+8896.4% 🔺)	3.490s (+1.5%)	1.833s	10	1.03x
▲ Vercel	Next.js (Turbopack)	1.785s (-4.5%)	3.025s (+6.5% 🔺)	0.011s (+160.5% 🔺)	3.640s (+6.9% 🔺)	1.855s	10	1.11x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

Summary

Fastest Framework by World

Winner determined by most benchmark wins

World	🥇 Fastest Framework	Wins
💻 Local	Express	12/12
🐘 Postgres	Express	7/12
▲ Vercel	Nitro	7/12

Fastest World by Framework

Winner determined by most benchmark wins

Framework	🥇 Fastest World	Wins
Express	💻 Local	6/12
Next.js (Turbopack)	🌐 Redis	9/12
Nitro	🐘 Postgres	6/12

Column Definitions

• Workflow Time: Runtime reported by workflow (completedAt - createdAt) - primary metric
• TTFB: Time to First Byte - time from workflow start until first stream byte received (stream benchmarks only)
• Slurp: Time from first byte to complete stream consumption (stream benchmarks only)
• Wall Time: Total testbench time (trigger workflow + poll for result)
• Overhead: Testbench overhead (Wall Time - Workflow Time)
• Samples: Number of benchmark iterations run
• vs Fastest: How much slower compared to the fastest configuration for this benchmark

Worlds:

• 💻 Local: In-memory filesystem world (local development)
• 🐘 Postgres: PostgreSQL database world (local development)
• ▲ Vercel: Vercel production/preview deployment
• 🌐 Turso: Community world (local development)
• 🌐 MongoDB: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Jazz: Community world (local development)

---

📋 View full workflow run

[pranaygp]

Suggested change

	'Run already finished during setup, skipping',
	'Run already finished during setup, skipping replay',

maybe this would be clearer to the user? (assuming the codepath is only for run replays)

[pranaygp]

or "skipping redundant workflow execution"

[pranaygp]

human: LGTM! great detailed state check and ty for posting that on github PR review comment

[VaguelySerious]

Noticed error logs from 409s falling through the queue, which are confusing for users. Log example:

Queue callback error: Error [WorkflowAPIError]: Cannot update workflow run wrun_01KKVYQ2TNA4PCEH9MA0SHN7DV with status 'completed'. Operation requires status 'running' or 'pending'.
at cP (.next/server/chunks/[root-of-the-server]__fd848f42._.js:63:41)
at async gA (.next/server/chunks/[root-of-the-server]__fd848f42._.js:67:3954)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:86:8)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:60:148157)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:60:146659)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:83:2206)
at async default (.next/server/chunks/[root-of-the-server]__fd848f42._.js:67:7813)
at async u7.processMessage (.next/server/chunks/[root-of-the-server]__fd848f42._.js:62:13992)
at async u7.consume (.next/server/chunks/[root-of-the-server]__fd848f42._.js:62:15197) {
cause: undefined,
status: 409,
code: undefined,
url: 'https://vercel-workflow.com/api/v1/runs/wrun_01KKVYQ2TNA4PCEH9MA0SHN7DV'
}

so I ensures we catch and gracefully return for these cases. I had Claude run an eval on whether this blocks any real use cases. Eval below:

---

Concurrent scenarios and why skipping is safe

Scenario 1: Two handlers race on run_started

Two queue messages for the same run arrive while it's still pending. Both handlers call world.runs.get(), see pending, and try to create a run_started event.

Handler A: runs.get() → pending → events.create(run_started) → ✅ succeeds → continues replay loop
Handler B: runs.get() → pending → events.create(run_started) → ❌ 409 → returns early

Why it's safe: Handler A transitions the run to running and enters the replay loop. The replay loop either completes the workflow or suspends it with queued continuations. Handler B's early exit is a no-op — Handler A guarantees progress.

Location:runtime.ts catch block after run_started event creation.

Scenario 2: Two handlers race on run_completed

Two concurrent replay loops both reach the end of the workflow and try to create run_completed.

Handler A: events.create(run_completed) → ✅ succeeds → returns
Handler B: events.create(run_completed) → ❌ 409 → logs info → returns

Why it's safe: The run is completed. Both handlers return. No continuation is needed.

Location:runtime.ts catch block after run_completed event creation.

Scenario 3: Two handlers race on run_failed

Same as Scenario 2 but for failure. Both handlers detect an error in user code and try to fail the run.

Handler A: events.create(run_failed) → ✅ succeeds → returns
Handler B: events.create(run_failed) → ❌ 409 → logs info → returns

Why it's safe: The run is failed. Both handlers return. No continuation is needed.

Location:runtime.ts catch block after run_failed event creation.

Scenario 4: Run completes while another handler creates hooks

Handler A completes the workflow. Handler B (from an earlier queue message) is still in the suspension handler creating hook events.

Handler A: events.create(run_completed) → ✅ run is now terminal
Handler B: events.create(hook_created) → ❌ 409 (run terminal) → logs info, continues
→ returns from handleSuspension with pendingSteps
→ tries to execute inline step
→ events.create(step_started) → ❌ 410 (run gone) → returns { type: 'gone' }
→ caller returns

Why it's safe: Handler A already completed the workflow. Handler B's hook creation fails, and if any steps were queued, the step executor receives 410 on step_started and exits gracefully. The 410 on step_started is the safety net — even if the suspension handler returns pending steps after a 409, the step executor won't make progress on a finished run.

Location:suspension-handler.ts catch blocks for hook_created and hook_disposed; step-executor.ts 410 handling on step_started.

Scenario 5: Two handlers race on wait_completed

During the replay loop, both handlers see an elapsed wait and try to complete it.

Handler A: events.create(wait_completed) → ✅ succeeds → adds event to cache → continues replay
Handler B: events.create(wait_completed) → ❌ 409 → continue (skip this wait) → continues replay

Why it's safe: Both handlers continue their replay loops. The event log is the same for both (the winning handler's wait_completed is visible to future reads). Subsequent replay from either handler converges to the same state because replay is deterministic and event-sourced.

Location:runtime.ts wait completion loop with continue on 409.

Scenario 6: Two handlers race on step_completed

Two handlers execute the same step concurrently (e.g., both received the step's queue message). Both finish execution and try to create step_completed.

Handler A: events.create(step_completed) → ✅ succeeds → queues workflow continuation → returns
Handler B: events.create(step_completed) → ❌ 409 → returns WITHOUT queuing continuation

Why it's safe: Handler A queues the workflow continuation. Handler B does not — this is correct because only one continuation should be queued per step completion. If both queued, there would be redundant replay (which is safe but wasteful).

Location:step-executor.ts catch on step_completed event creation.

Scenario 7: Two handlers race on step_started

Two handlers try to start the same step. Handler A wins; Handler B gets 409 because the step transitioned to a terminal state.

Handler A: events.create(step_started) → ✅ succeeds → executes step → queues continuation
Handler B: events.create(step_started) → ❌ 409 (step terminal) → returns { type: 'skipped' }
→ caller queues workflow continuation (runtime.ts replay loop)

Why it's safe: Both handlers ensure workflow continuation — Handler A via step completion, Handler B via the replay loop re-queuing the workflow. The redundant replay is safe because event-sourced replay is convergent.

Location:step-executor.ts 409 handling on step_started.

Scenario 8: step_created/wait_created duplicate

During suspension handling, two concurrent replays both try to create the same step or wait event.

Handler A: events.create(step_created) → ✅ succeeds
Handler B: events.create(step_created) → ❌ 409 → logs info → continues

Why it's safe: The step/wait entity already exists. The suspension handler continues processing other items. The step will be executed by whichever handler picks it up from the queue.

Location:suspension-handler.ts catch blocks for step_created and wait_created.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
example-nextjs-workflow-turbopack	Ready	Preview, Comment	Mar 17, 2026 5:00pm
example-nextjs-workflow-webpack	Ready	Preview, Comment	Mar 17, 2026 5:00pm
example-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-astro-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-express-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-fastify-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-hono-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-nitro-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-nuxt-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-sveltekit-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-vite-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workflow-docs	Ready	Preview, Comment, Open in v0	Mar 17, 2026 5:00pm
workflow-nest	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workflow-swc-playground	Ready	Preview, Comment	Mar 17, 2026 5:00pm

[changeset-bot]

🦋 Changeset detected

Latest commit: 9668d09

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 16 packages

Name	Type
@workflow/core	Patch
@workflow/builders	Patch
@workflow/cli	Patch
@workflow/next	Patch
@workflow/nitro	Patch
@workflow/vitest	Patch
@workflow/web-shared	Patch
workflow	Patch
@workflow/world-testing	Patch
@workflow/astro	Patch
@workflow/nest	Patch
@workflow/rollup	Patch
@workflow/sveltekit	Patch
@workflow/vite	Patch
@workflow/nuxt	Patch
@workflow/ai	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

🧪 E2E Test Results

❌ Some tests failed

Summary

	Passed	Failed	Skipped	Total
✅ ▲ Vercel Production	747	0	67	814
✅ 💻 Local Development	770	0	118	888
✅ 📦 Local Production	770	0	118	888
✅ 🐘 Local Postgres	770	0	118	888
✅ 🪟 Windows	71	0	3	74
❌ 🌍 Community Worlds	116	55	15	186
✅ 📋 Other	195	0	27	222
Total	3439	55	466	3960

❌ Failed Tests

🌍 Community Worlds (55 failed)

mongodb (3 failed):

• hookWorkflow is not resumable via public webhook endpoint
• webhookWorkflow
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously

redis (2 failed):

• hookWorkflow is not resumable via public webhook endpoint
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously

turso (50 failed):

• addTenWorkflow
• addTenWorkflow
• wellKnownAgentWorkflow (.well-known/agent)
• should work with react rendering in step
• promiseAllWorkflow
• promiseRaceWorkflow
• promiseAnyWorkflow
• importedStepOnlyWorkflow
• hookWorkflow
• hookWorkflow is not resumable via public webhook endpoint
• webhookWorkflow
• sleepingWorkflow
• parallelSleepWorkflow
• nullByteWorkflow
• workflowAndStepMetadataWorkflow
• fetchWorkflow
• promiseRaceStressTestWorkflow
• error handling error propagation workflow errors nested function calls preserve message and stack trace
• error handling error propagation workflow errors cross-file imports preserve message and stack trace
• error handling error propagation step errors basic step error preserves message and stack trace
• error handling error propagation step errors cross-file step error preserves message and function names in stack
• error handling retry behavior regular Error retries until success
• error handling retry behavior FatalError fails immediately without retries
• error handling retry behavior RetryableError respects custom retryAfter delay
• error handling retry behavior maxRetries=0 disables retries
• error handling catchability FatalError can be caught and detected with FatalError.is()
• hookCleanupTestWorkflow - hook token reuse after workflow completion
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously
• hookDisposeTestWorkflow - hook token reuse after explicit disposal while workflow still running
• stepFunctionPassingWorkflow - step function references can be passed as arguments (without closure vars)
• stepFunctionWithClosureWorkflow - step function with closure variables passed as argument
• closureVariableWorkflow - nested step functions with closure variables
• spawnWorkflowFromStepWorkflow - spawning a child workflow using start() inside a step
• health check (queue-based) - workflow and step endpoints respond to health check messages
• pathsAliasWorkflow - TypeScript path aliases resolve correctly
• Calculator.calculate - static workflow method using static step methods from another class
• AllInOneService.processNumber - static workflow method using sibling static step methods
• ChainableService.processWithThis - static step methods using this to reference the class
• thisSerializationWorkflow - step function invoked with .call() and .apply()
• customSerializationWorkflow - custom class serialization with WORKFLOW_SERIALIZE/WORKFLOW_DESERIALIZE
• instanceMethodStepWorkflow - instance methods with "use step" directive
• crossContextSerdeWorkflow - classes defined in step code are deserializable in workflow context
• stepFunctionAsStartArgWorkflow - step function reference passed as start() argument
• cancelRun - cancelling a running workflow
• cancelRun via CLI - cancelling a running workflow
• pages router addTenWorkflow via pages router
• pages router promiseAllWorkflow via pages router
• pages router sleepingWorkflow via pages router
• hookWithSleepWorkflow - hook payloads delivered correctly with concurrent sleep
• sleepWithSequentialStepsWorkflow - sequential steps work with concurrent sleep (control)

Details by Category

✅ ▲ Vercel Production

App	Passed	Failed	Skipped
✅ astro	67	0	7
✅ example	67	0	7
✅ express	67	0	7
✅ fastify	67	0	7
✅ hono	67	0	7
✅ nextjs-turbopack	72	0	2
✅ nextjs-webpack	72	0	2
✅ nitro	67	0	7
✅ nuxt	67	0	7
✅ sveltekit	67	0	7
✅ vite	67	0	7

✅ 💻 Local Development

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 📦 Local Production

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 🐘 Local Postgres

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 🪟 Windows

App	Passed	Failed	Skipped
✅ nextjs-turbopack	71	0	3

❌ 🌍 Community Worlds

App	Passed	Failed	Skipped
✅ mongodb-dev	3	0	2
❌ mongodb	51	3	3
✅ redis-dev	3	0	2
❌ redis	52	2	3
✅ turso-dev	3	0	2
❌ turso	4	50	3

✅ 📋 Other

App	Passed	Failed	Skipped
✅ e2e-local-dev-nest-stable	65	0	9
✅ e2e-local-postgres-nest-stable	65	0	9
✅ e2e-local-prod-nest-stable	65	0	9

---

📋 View full workflow run

[github-actions]

📊 Benchmark Results

📈 Comparing against baseline from main branch. Green 🟢 = faster, Red 🔺 = slower.

workflow with no steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	0.038s (-15.7% 🟢)	1.005s (~)	0.968s	10	1.00x
💻 Local	Nitro	0.046s (+6.7% 🔺)	1.006s (~)	0.959s	10	1.23x
💻 Local	Next.js (Turbopack)	0.051s	1.006s	0.955s	10	1.36x
🌐 Redis	Next.js (Turbopack)	0.054s	1.005s	0.951s	10	1.44x
🐘 Postgres	Nitro	0.059s (-4.7%)	1.011s (~)	0.952s	10	1.56x
🐘 Postgres	Express	0.059s (-14.8% 🟢)	1.011s (~)	0.952s	10	1.57x
🌐 MongoDB	Next.js (Turbopack)	0.079s	1.008s	0.929s	10	2.11x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	0.465s (-33.5% 🟢)	2.036s (-22.7% 🟢)	1.571s	10	1.00x
▲ Vercel	Express	0.504s (-20.0% 🟢)	2.286s (-14.4% 🟢)	1.781s	10	1.08x
▲ Vercel	Next.js (Turbopack)	0.511s (-17.6% 🟢)	2.559s (+13.2% 🔺)	2.049s	10	1.10x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

workflow with 1 step

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.093s (-3.3%)	2.006s (~)	0.912s	10	1.00x
🌐 Redis	Next.js (Turbopack)	1.122s	2.006s	0.885s	10	1.03x
💻 Local	Nitro	1.130s (~)	2.006s (~)	0.877s	10	1.03x
💻 Local	Next.js (Turbopack)	1.130s	2.006s	0.876s	10	1.03x
🐘 Postgres	Express	1.153s (+0.7%)	2.014s (~)	0.860s	10	1.05x
🐘 Postgres	Nitro	1.158s (+1.1%)	2.012s (~)	0.855s	10	1.06x
🌐 MongoDB	Next.js (Turbopack)	1.307s	2.008s	0.700s	10	1.20x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.040s (-3.0%)	3.224s (-4.4%)	1.184s	10	1.00x
▲ Vercel	Express	2.053s (+0.9%)	3.735s (+12.4% 🔺)	1.682s	10	1.01x
▲ Vercel	Next.js (Turbopack)	2.168s (-2.0%)	3.924s (+13.8% 🔺)	1.756s	10	1.06x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

workflow with 10 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	10.615s (-2.8%)	11.023s (~)	0.408s	3	1.00x
🌐 Redis	Next.js (Turbopack)	10.784s	11.023s	0.239s	3	1.02x
💻 Local	Next.js (Turbopack)	10.812s	11.023s	0.211s	3	1.02x
🐘 Postgres	Nitro	10.927s (~)	11.042s (~)	0.115s	3	1.03x
🐘 Postgres	Express	10.970s (~)	11.375s (~)	0.405s	3	1.03x
💻 Local	Nitro	10.976s (+0.8%)	11.024s (~)	0.048s	3	1.03x
🌐 MongoDB	Next.js (Turbopack)	12.203s	13.015s	0.812s	3	1.15x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	16.779s (-7.4% 🟢)	18.655s (-5.6% 🟢)	1.876s	2	1.00x
▲ Vercel	Nitro	16.868s (+1.3%)	18.197s (-1.3%)	1.329s	2	1.01x
▲ Vercel	Next.js (Turbopack)	17.826s (+0.9%)	19.468s (+1.2%)	1.641s	2	1.06x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

workflow with 25 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	26.743s (-3.1%)	27.051s (-3.6%)	0.308s	3	1.00x
🌐 Redis	Next.js (Turbopack)	26.760s	27.049s	0.289s	3	1.00x
💻 Local	Next.js (Turbopack)	27.128s	28.053s	0.925s	3	1.01x
🐘 Postgres	Express	27.240s (~)	28.066s (~)	0.826s	3	1.02x
🐘 Postgres	Nitro	27.277s (~)	28.065s (~)	0.788s	3	1.02x
💻 Local	Nitro	27.583s (~)	28.053s (~)	0.470s	3	1.03x
🌐 MongoDB	Next.js (Turbopack)	30.482s	31.045s	0.564s	2	1.14x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	46.028s (+2.9%)	47.737s (+4.3%)	1.709s	2	1.00x
▲ Vercel	Next.js (Turbopack)	50.377s (+6.6% 🔺)	52.253s (+8.0% 🔺)	1.877s	2	1.09x
▲ Vercel	Express	53.606s (+21.0% 🔺)	55.319s (+22.1% 🔺)	1.713s	2	1.16x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

workflow with 50 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	53.541s	54.093s	0.552s	2	1.00x
🐘 Postgres	Express	54.300s (~)	55.094s (~)	0.793s	2	1.01x
🐘 Postgres	Nitro	54.476s (~)	55.105s (~)	0.629s	2	1.02x
💻 Local	Express	54.851s (-3.3%)	55.101s (-3.5%)	0.250s	2	1.02x
💻 Local	Next.js (Turbopack)	55.951s	56.100s	0.149s	2	1.05x
💻 Local	Nitro	56.657s (~)	57.106s (~)	0.450s	2	1.06x
🌐 MongoDB	Next.js (Turbopack)	60.675s	61.065s	0.390s	2	1.13x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	94.771s (-4.4%)	96.975s (-4.2%)	2.204s	1	1.00x
▲ Vercel	Nitro	96.605s (~)	98.384s (~)	1.779s	1	1.02x
▲ Vercel	Next.js (Turbopack)	98.753s (-3.9%)	100.796s (-2.8%)	2.043s	1	1.04x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

Promise.all with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	1.345s	2.006s	0.662s	15	1.00x
🐘 Postgres	Express	1.366s (-4.0%)	2.011s (~)	0.645s	15	1.02x
💻 Local	Express	1.467s (-5.2% 🟢)	2.006s (~)	0.538s	15	1.09x
💻 Local	Nitro	1.498s (-1.1%)	2.005s (~)	0.507s	15	1.11x
💻 Local	Next.js (Turbopack)	1.547s	2.006s	0.459s	15	1.15x
🌐 MongoDB	Next.js (Turbopack)	2.146s	3.009s	0.863s	10	1.60x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-
🐘 Postgres	Nitro	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.312s (-4.2%)	3.407s (-4.8%)	1.095s	9	1.00x
▲ Vercel	Next.js (Turbopack)	2.641s (-1.3%)	4.038s (+7.5% 🔺)	1.397s	8	1.14x
▲ Vercel	Express	2.648s (+5.0%)	4.138s (+13.0% 🔺)	1.490s	8	1.15x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Promise.all with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	2.567s	3.008s	0.440s	10	1.00x
🐘 Postgres	Nitro	2.595s (~)	3.014s (~)	0.419s	10	1.01x
💻 Local	Express	2.600s (-14.1% 🟢)	3.007s (-15.6% 🟢)	0.407s	10	1.01x
🐘 Postgres	Express	2.615s (~)	3.015s (~)	0.399s	10	1.02x
💻 Local	Next.js (Turbopack)	3.041s	3.760s	0.718s	8	1.18x
💻 Local	Nitro	3.121s (+7.5% 🔺)	3.565s (+11.1% 🔺)	0.444s	9	1.22x
🌐 MongoDB	Next.js (Turbopack)	4.747s	5.179s	0.432s	6	1.85x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Next.js (Turbopack)	2.799s (+7.7% 🔺)	4.348s (+21.1% 🔺)	1.550s	8	1.00x
▲ Vercel	Nitro	3.520s (+30.2% 🔺)	4.708s (+23.4% 🔺)	1.188s	7	1.26x
▲ Vercel	Express	3.663s (+47.9% 🔺)	5.183s (+40.8% 🔺)	1.520s	8	1.31x

🔍 Observability: Next.js (Turbopack) | Nitro | Express

Promise.all with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	4.027s (+1.4%)	4.588s (+3.2%)	0.561s	7	1.00x
🐘 Postgres	Nitro	4.032s (+0.7%)	4.590s (+3.1%)	0.558s	7	1.00x
🌐 Redis	Next.js (Turbopack)	4.081s	5.012s	0.931s	6	1.01x
💻 Local	Express	6.787s (-15.7% 🟢)	7.014s (-20.1% 🟢)	0.227s	5	1.69x
💻 Local	Next.js (Turbopack)	7.888s	8.517s	0.629s	4	1.96x
💻 Local	Nitro	8.207s (-1.6%)	8.521s (-5.5% 🟢)	0.314s	4	2.04x
🌐 MongoDB	Next.js (Turbopack)	10.067s	10.684s	0.617s	3	2.50x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Next.js (Turbopack)	5.520s (+44.3% 🔺)	7.146s (+37.1% 🔺)	1.626s	5	1.00x
▲ Vercel	Nitro	7.009s (+153.3% 🔺)	8.352s (+119.6% 🔺)	1.343s	4	1.27x
▲ Vercel	Express	10.371s (+264.0% 🔺)	12.346s (+209.2% 🔺)	1.975s	3	1.88x

🔍 Observability: Next.js (Turbopack) | Nitro | Express

Promise.race with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	1.303s	2.006s	0.703s	15	1.00x
🐘 Postgres	Nitro	1.422s (-1.3%)	2.011s (~)	0.589s	15	1.09x
🐘 Postgres	Express	1.442s (-1.5%)	2.011s (-3.2%)	0.570s	15	1.11x
💻 Local	Express	1.477s (-2.9%)	2.005s (~)	0.528s	15	1.13x
💻 Local	Nitro	1.523s (-1.1%)	2.006s (~)	0.483s	15	1.17x
💻 Local	Next.js (Turbopack)	1.564s	2.073s	0.509s	15	1.20x
🌐 MongoDB	Next.js (Turbopack)	2.174s	3.009s	0.835s	10	1.67x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.131s (-0.9%)	3.290s (-1.5%)	1.159s	10	1.00x
▲ Vercel	Next.js (Turbopack)	2.268s (-27.5% 🟢)	3.823s (-24.1% 🟢)	1.555s	8	1.06x
▲ Vercel	Express	2.591s (+17.2% 🔺)	4.029s (+17.2% 🔺)	1.438s	8	1.22x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Promise.race with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	2.526s (-3.2%)	3.012s (~)	0.486s	10	1.00x
🌐 Redis	Next.js (Turbopack)	2.532s	3.008s	0.477s	10	1.00x
🐘 Postgres	Nitro	2.730s (+7.4% 🔺)	3.015s (~)	0.285s	10	1.08x
💻 Local	Express	2.797s (-9.8% 🟢)	3.108s (-17.3% 🟢)	0.311s	10	1.11x
💻 Local	Next.js (Turbopack)	3.030s	3.760s	0.729s	8	1.20x
💻 Local	Nitro	3.124s (-1.8%)	3.886s (~)	0.762s	8	1.24x
🌐 MongoDB	Next.js (Turbopack)	4.728s	5.177s	0.449s	6	1.87x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.270s (-29.0% 🟢)	3.269s (-37.5% 🟢)	0.998s	10	1.00x
▲ Vercel	Express	2.378s (-32.1% 🟢)	3.827s (-17.2% 🟢)	1.449s	8	1.05x
▲ Vercel	Next.js (Turbopack)	3.093s (-5.5% 🟢)	4.610s (~)	1.517s	7	1.36x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

Promise.race with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	3.975s (-3.5%)	4.590s (~)	0.615s	7	1.00x
🐘 Postgres	Express	4.039s (+2.3%)	4.590s (+3.2%)	0.551s	7	1.02x
🌐 Redis	Next.js (Turbopack)	4.198s	4.725s	0.527s	7	1.06x
💻 Local	Express	7.733s (-14.1% 🟢)	8.267s (-10.8% 🟢)	0.533s	4	1.95x
💻 Local	Next.js (Turbopack)	8.255s	8.766s	0.512s	4	2.08x
💻 Local	Nitro	8.519s (+0.8%)	9.275s (+2.8%)	0.756s	4	2.14x
🌐 MongoDB	Next.js (Turbopack)	10.032s	10.680s	0.648s	3	2.52x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.724s (-27.3% 🟢)	3.647s (-28.7% 🟢)	0.923s	9	1.00x
▲ Vercel	Next.js (Turbopack)	3.571s (-3.3%)	5.369s (+9.7% 🔺)	1.798s	6	1.31x
▲ Vercel	Express	3.789s (+1.4%)	5.100s (-0.7%)	1.311s	6	1.39x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Stream Benchmarks(includes TTFB metrics)

workflow with stream

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	0.138s (-32.2% 🟢)	1.003s (~)	0.009s (-21.6% 🟢)	1.015s (~)	0.877s	10	1.00x
💻 Local	Next.js (Turbopack)	0.170s	1.002s	0.011s	1.017s	0.847s	10	1.24x
🌐 Redis	Next.js (Turbopack)	0.182s	1.000s	0.002s	1.007s	0.825s	10	1.33x
💻 Local	Nitro	0.212s (+7.3% 🔺)	1.003s (~)	0.011s (-2.7%)	1.017s (~)	0.804s	10	1.54x
🐘 Postgres	Express	0.215s (+2.8%)	0.996s (~)	0.002s (+15.4% 🔺)	1.013s (~)	0.797s	10	1.57x
🐘 Postgres	Nitro	0.246s (+9.4% 🔺)	0.993s (~)	0.002s (+41.7% 🔺)	1.014s (~)	0.767s	10	1.79x
🌐 MongoDB	Next.js (Turbopack)	0.474s	0.979s	0.002s	1.009s	0.534s	10	3.45x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	1.606s (+1.9%)	2.405s (-8.7% 🟢)	0.006s (-26.2% 🟢)	3.035s (-2.2%)	1.429s	10	1.00x
▲ Vercel	Nitro	1.658s (+1.2%)	2.506s (-14.3% 🟢)	0.504s (+8896.4% 🔺)	3.490s (+1.5%)	1.833s	10	1.03x
▲ Vercel	Next.js (Turbopack)	1.785s (-4.5%)	3.025s (+6.5% 🔺)	0.011s (+160.5% 🔺)	3.640s (+6.9% 🔺)	1.855s	10	1.11x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

Summary

Fastest Framework by World

Winner determined by most benchmark wins

World	🥇 Fastest Framework	Wins
💻 Local	Express	12/12
🐘 Postgres	Express	7/12
▲ Vercel	Nitro	7/12

Fastest World by Framework

Winner determined by most benchmark wins

Framework	🥇 Fastest World	Wins
Express	💻 Local	6/12
Next.js (Turbopack)	🌐 Redis	9/12
Nitro	🐘 Postgres	6/12

Column Definitions

• Workflow Time: Runtime reported by workflow (completedAt - createdAt) - primary metric
• TTFB: Time to First Byte - time from workflow start until first stream byte received (stream benchmarks only)
• Slurp: Time from first byte to complete stream consumption (stream benchmarks only)
• Wall Time: Total testbench time (trigger workflow + poll for result)
• Overhead: Testbench overhead (Wall Time - Workflow Time)
• Samples: Number of benchmark iterations run
• vs Fastest: How much slower compared to the fastest configuration for this benchmark

Worlds:

• 💻 Local: In-memory filesystem world (local development)
• 🐘 Postgres: PostgreSQL database world (local development)
• ▲ Vercel: Vercel production/preview deployment
• 🌐 Turso: Community world (local development)
• 🌐 MongoDB: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Jazz: Community world (local development)

---

📋 View full workflow run

[pranaygp]

Suggested change

	'Run already finished during setup, skipping',
	'Run already finished during setup, skipping replay',

maybe this would be clearer to the user? (assuming the codepath is only for run replays)

[pranaygp]

or "skipping redundant workflow execution"

[pranaygp]

human: LGTM! great detailed state check and ty for posting that on github PR review comment

[VaguelySerious]

Noticed error logs from 409s falling through the queue, which are confusing for users. Log example:

Queue callback error: Error [WorkflowAPIError]: Cannot update workflow run wrun_01KKVYQ2TNA4PCEH9MA0SHN7DV with status 'completed'. Operation requires status 'running' or 'pending'.
at cP (.next/server/chunks/[root-of-the-server]__fd848f42._.js:63:41)
at async gA (.next/server/chunks/[root-of-the-server]__fd848f42._.js:67:3954)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:86:8)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:60:148157)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:60:146659)
at async (.next/server/chunks/[root-of-the-server]__fd848f42._.js:83:2206)
at async default (.next/server/chunks/[root-of-the-server]__fd848f42._.js:67:7813)
at async u7.processMessage (.next/server/chunks/[root-of-the-server]__fd848f42._.js:62:13992)
at async u7.consume (.next/server/chunks/[root-of-the-server]__fd848f42._.js:62:15197) {
cause: undefined,
status: 409,
code: undefined,
url: 'https://vercel-workflow.com/api/v1/runs/wrun_01KKVYQ2TNA4PCEH9MA0SHN7DV'
}

so I ensures we catch and gracefully return for these cases. I had Claude run an eval on whether this blocks any real use cases. Eval below:

---

Concurrent scenarios and why skipping is safe

Scenario 1: Two handlers race on run_started

Two queue messages for the same run arrive while it's still pending. Both handlers call world.runs.get(), see pending, and try to create a run_started event.

Handler A: runs.get() → pending → events.create(run_started) → ✅ succeeds → continues replay loop
Handler B: runs.get() → pending → events.create(run_started) → ❌ 409 → returns early

Why it's safe: Handler A transitions the run to running and enters the replay loop. The replay loop either completes the workflow or suspends it with queued continuations. Handler B's early exit is a no-op — Handler A guarantees progress.

Location:runtime.ts catch block after run_started event creation.

Scenario 2: Two handlers race on run_completed

Two concurrent replay loops both reach the end of the workflow and try to create run_completed.

Handler A: events.create(run_completed) → ✅ succeeds → returns
Handler B: events.create(run_completed) → ❌ 409 → logs info → returns

Why it's safe: The run is completed. Both handlers return. No continuation is needed.

Location:runtime.ts catch block after run_completed event creation.

Scenario 3: Two handlers race on run_failed

Same as Scenario 2 but for failure. Both handlers detect an error in user code and try to fail the run.

Handler A: events.create(run_failed) → ✅ succeeds → returns
Handler B: events.create(run_failed) → ❌ 409 → logs info → returns

Why it's safe: The run is failed. Both handlers return. No continuation is needed.

Location:runtime.ts catch block after run_failed event creation.

Scenario 4: Run completes while another handler creates hooks

Handler A completes the workflow. Handler B (from an earlier queue message) is still in the suspension handler creating hook events.

Handler A: events.create(run_completed) → ✅ run is now terminal
Handler B: events.create(hook_created) → ❌ 409 (run terminal) → logs info, continues
→ returns from handleSuspension with pendingSteps
→ tries to execute inline step
→ events.create(step_started) → ❌ 410 (run gone) → returns { type: 'gone' }
→ caller returns

Why it's safe: Handler A already completed the workflow. Handler B's hook creation fails, and if any steps were queued, the step executor receives 410 on step_started and exits gracefully. The 410 on step_started is the safety net — even if the suspension handler returns pending steps after a 409, the step executor won't make progress on a finished run.

Location:suspension-handler.ts catch blocks for hook_created and hook_disposed; step-executor.ts 410 handling on step_started.

Scenario 5: Two handlers race on wait_completed

During the replay loop, both handlers see an elapsed wait and try to complete it.

Handler A: events.create(wait_completed) → ✅ succeeds → adds event to cache → continues replay
Handler B: events.create(wait_completed) → ❌ 409 → continue (skip this wait) → continues replay

Why it's safe: Both handlers continue their replay loops. The event log is the same for both (the winning handler's wait_completed is visible to future reads). Subsequent replay from either handler converges to the same state because replay is deterministic and event-sourced.

Location:runtime.ts wait completion loop with continue on 409.

Scenario 6: Two handlers race on step_completed

Two handlers execute the same step concurrently (e.g., both received the step's queue message). Both finish execution and try to create step_completed.

Handler A: events.create(step_completed) → ✅ succeeds → queues workflow continuation → returns
Handler B: events.create(step_completed) → ❌ 409 → returns WITHOUT queuing continuation

Why it's safe: Handler A queues the workflow continuation. Handler B does not — this is correct because only one continuation should be queued per step completion. If both queued, there would be redundant replay (which is safe but wasteful).

Location:step-executor.ts catch on step_completed event creation.

Scenario 7: Two handlers race on step_started

Two handlers try to start the same step. Handler A wins; Handler B gets 409 because the step transitioned to a terminal state.

Handler A: events.create(step_started) → ✅ succeeds → executes step → queues continuation
Handler B: events.create(step_started) → ❌ 409 (step terminal) → returns { type: 'skipped' }
→ caller queues workflow continuation (runtime.ts replay loop)

Why it's safe: Both handlers ensure workflow continuation — Handler A via step completion, Handler B via the replay loop re-queuing the workflow. The redundant replay is safe because event-sourced replay is convergent.

Location:step-executor.ts 409 handling on step_started.

Scenario 8: step_created/wait_created duplicate

During suspension handling, two concurrent replays both try to create the same step or wait event.

Handler A: events.create(step_created) → ✅ succeeds
Handler B: events.create(step_created) → ❌ 409 → logs info → continues

Why it's safe: The step/wait entity already exists. The suspension handler continues processing other items. The step will be executed by whichever handler picks it up from the queue.

Location:suspension-handler.ts catch blocks for step_created and wait_created.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
example-nextjs-workflow-turbopack	Ready	Preview, Comment	Mar 17, 2026 5:00pm
example-nextjs-workflow-webpack	Ready	Preview, Comment	Mar 17, 2026 5:00pm
example-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-astro-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-express-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-fastify-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-hono-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-nitro-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-nuxt-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-sveltekit-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workbench-vite-workflow	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workflow-docs	Ready	Preview, Comment, Open in v0	Mar 17, 2026 5:00pm
workflow-nest	Ready	Preview, Comment	Mar 17, 2026 5:00pm
workflow-swc-playground	Ready	Preview, Comment	Mar 17, 2026 5:00pm

[changeset-bot]

🦋 Changeset detected

Latest commit: 9668d09

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 16 packages

Name	Type
@workflow/core	Patch
@workflow/builders	Patch
@workflow/cli	Patch
@workflow/next	Patch
@workflow/nitro	Patch
@workflow/vitest	Patch
@workflow/web-shared	Patch
workflow	Patch
@workflow/world-testing	Patch
@workflow/astro	Patch
@workflow/nest	Patch
@workflow/rollup	Patch
@workflow/sveltekit	Patch
@workflow/vite	Patch
@workflow/nuxt	Patch
@workflow/ai	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

🧪 E2E Test Results

❌ Some tests failed

Summary

	Passed	Failed	Skipped	Total
✅ ▲ Vercel Production	747	0	67	814
✅ 💻 Local Development	770	0	118	888
✅ 📦 Local Production	770	0	118	888
✅ 🐘 Local Postgres	770	0	118	888
✅ 🪟 Windows	71	0	3	74
❌ 🌍 Community Worlds	116	55	15	186
✅ 📋 Other	195	0	27	222
Total	3439	55	466	3960

❌ Failed Tests

🌍 Community Worlds (55 failed)

mongodb (3 failed):

• hookWorkflow is not resumable via public webhook endpoint
• webhookWorkflow
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously

redis (2 failed):

• hookWorkflow is not resumable via public webhook endpoint
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously

turso (50 failed):

• addTenWorkflow
• addTenWorkflow
• wellKnownAgentWorkflow (.well-known/agent)
• should work with react rendering in step
• promiseAllWorkflow
• promiseRaceWorkflow
• promiseAnyWorkflow
• importedStepOnlyWorkflow
• hookWorkflow
• hookWorkflow is not resumable via public webhook endpoint
• webhookWorkflow
• sleepingWorkflow
• parallelSleepWorkflow
• nullByteWorkflow
• workflowAndStepMetadataWorkflow
• fetchWorkflow
• promiseRaceStressTestWorkflow
• error handling error propagation workflow errors nested function calls preserve message and stack trace
• error handling error propagation workflow errors cross-file imports preserve message and stack trace
• error handling error propagation step errors basic step error preserves message and stack trace
• error handling error propagation step errors cross-file step error preserves message and function names in stack
• error handling retry behavior regular Error retries until success
• error handling retry behavior FatalError fails immediately without retries
• error handling retry behavior RetryableError respects custom retryAfter delay
• error handling retry behavior maxRetries=0 disables retries
• error handling catchability FatalError can be caught and detected with FatalError.is()
• hookCleanupTestWorkflow - hook token reuse after workflow completion
• concurrent hook token conflict - two workflows cannot use the same hook token simultaneously
• hookDisposeTestWorkflow - hook token reuse after explicit disposal while workflow still running
• stepFunctionPassingWorkflow - step function references can be passed as arguments (without closure vars)
• stepFunctionWithClosureWorkflow - step function with closure variables passed as argument
• closureVariableWorkflow - nested step functions with closure variables
• spawnWorkflowFromStepWorkflow - spawning a child workflow using start() inside a step
• health check (queue-based) - workflow and step endpoints respond to health check messages
• pathsAliasWorkflow - TypeScript path aliases resolve correctly
• Calculator.calculate - static workflow method using static step methods from another class
• AllInOneService.processNumber - static workflow method using sibling static step methods
• ChainableService.processWithThis - static step methods using this to reference the class
• thisSerializationWorkflow - step function invoked with .call() and .apply()
• customSerializationWorkflow - custom class serialization with WORKFLOW_SERIALIZE/WORKFLOW_DESERIALIZE
• instanceMethodStepWorkflow - instance methods with "use step" directive
• crossContextSerdeWorkflow - classes defined in step code are deserializable in workflow context
• stepFunctionAsStartArgWorkflow - step function reference passed as start() argument
• cancelRun - cancelling a running workflow
• cancelRun via CLI - cancelling a running workflow
• pages router addTenWorkflow via pages router
• pages router promiseAllWorkflow via pages router
• pages router sleepingWorkflow via pages router
• hookWithSleepWorkflow - hook payloads delivered correctly with concurrent sleep
• sleepWithSequentialStepsWorkflow - sequential steps work with concurrent sleep (control)

Details by Category

✅ ▲ Vercel Production

App	Passed	Failed	Skipped
✅ astro	67	0	7
✅ example	67	0	7
✅ express	67	0	7
✅ fastify	67	0	7
✅ hono	67	0	7
✅ nextjs-turbopack	72	0	2
✅ nextjs-webpack	72	0	2
✅ nitro	67	0	7
✅ nuxt	67	0	7
✅ sveltekit	67	0	7
✅ vite	67	0	7

✅ 💻 Local Development

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 📦 Local Production

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 🐘 Local Postgres

App	Passed	Failed	Skipped
✅ astro-stable	65	0	9
✅ express-stable	65	0	9
✅ fastify-stable	65	0	9
✅ hono-stable	65	0	9
✅ nextjs-turbopack-canary	54	0	20
✅ nextjs-turbopack-stable	71	0	3
✅ nextjs-webpack-canary	54	0	20
✅ nextjs-webpack-stable	71	0	3
✅ nitro-stable	65	0	9
✅ nuxt-stable	65	0	9
✅ sveltekit-stable	65	0	9
✅ vite-stable	65	0	9

✅ 🪟 Windows

App	Passed	Failed	Skipped
✅ nextjs-turbopack	71	0	3

❌ 🌍 Community Worlds

App	Passed	Failed	Skipped
✅ mongodb-dev	3	0	2
❌ mongodb	51	3	3
✅ redis-dev	3	0	2
❌ redis	52	2	3
✅ turso-dev	3	0	2
❌ turso	4	50	3

✅ 📋 Other

App	Passed	Failed	Skipped
✅ e2e-local-dev-nest-stable	65	0	9
✅ e2e-local-postgres-nest-stable	65	0	9
✅ e2e-local-prod-nest-stable	65	0	9

---

📋 View full workflow run

[github-actions]

📊 Benchmark Results

📈 Comparing against baseline from main branch. Green 🟢 = faster, Red 🔺 = slower.

workflow with no steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	0.038s (-15.7% 🟢)	1.005s (~)	0.968s	10	1.00x
💻 Local	Nitro	0.046s (+6.7% 🔺)	1.006s (~)	0.959s	10	1.23x
💻 Local	Next.js (Turbopack)	0.051s	1.006s	0.955s	10	1.36x
🌐 Redis	Next.js (Turbopack)	0.054s	1.005s	0.951s	10	1.44x
🐘 Postgres	Nitro	0.059s (-4.7%)	1.011s (~)	0.952s	10	1.56x
🐘 Postgres	Express	0.059s (-14.8% 🟢)	1.011s (~)	0.952s	10	1.57x
🌐 MongoDB	Next.js (Turbopack)	0.079s	1.008s	0.929s	10	2.11x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	0.465s (-33.5% 🟢)	2.036s (-22.7% 🟢)	1.571s	10	1.00x
▲ Vercel	Express	0.504s (-20.0% 🟢)	2.286s (-14.4% 🟢)	1.781s	10	1.08x
▲ Vercel	Next.js (Turbopack)	0.511s (-17.6% 🟢)	2.559s (+13.2% 🔺)	2.049s	10	1.10x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

workflow with 1 step

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	1.093s (-3.3%)	2.006s (~)	0.912s	10	1.00x
🌐 Redis	Next.js (Turbopack)	1.122s	2.006s	0.885s	10	1.03x
💻 Local	Nitro	1.130s (~)	2.006s (~)	0.877s	10	1.03x
💻 Local	Next.js (Turbopack)	1.130s	2.006s	0.876s	10	1.03x
🐘 Postgres	Express	1.153s (+0.7%)	2.014s (~)	0.860s	10	1.05x
🐘 Postgres	Nitro	1.158s (+1.1%)	2.012s (~)	0.855s	10	1.06x
🌐 MongoDB	Next.js (Turbopack)	1.307s	2.008s	0.700s	10	1.20x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.040s (-3.0%)	3.224s (-4.4%)	1.184s	10	1.00x
▲ Vercel	Express	2.053s (+0.9%)	3.735s (+12.4% 🔺)	1.682s	10	1.01x
▲ Vercel	Next.js (Turbopack)	2.168s (-2.0%)	3.924s (+13.8% 🔺)	1.756s	10	1.06x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

workflow with 10 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	10.615s (-2.8%)	11.023s (~)	0.408s	3	1.00x
🌐 Redis	Next.js (Turbopack)	10.784s	11.023s	0.239s	3	1.02x
💻 Local	Next.js (Turbopack)	10.812s	11.023s	0.211s	3	1.02x
🐘 Postgres	Nitro	10.927s (~)	11.042s (~)	0.115s	3	1.03x
🐘 Postgres	Express	10.970s (~)	11.375s (~)	0.405s	3	1.03x
💻 Local	Nitro	10.976s (+0.8%)	11.024s (~)	0.048s	3	1.03x
🌐 MongoDB	Next.js (Turbopack)	12.203s	13.015s	0.812s	3	1.15x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	16.779s (-7.4% 🟢)	18.655s (-5.6% 🟢)	1.876s	2	1.00x
▲ Vercel	Nitro	16.868s (+1.3%)	18.197s (-1.3%)	1.329s	2	1.01x
▲ Vercel	Next.js (Turbopack)	17.826s (+0.9%)	19.468s (+1.2%)	1.641s	2	1.06x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

workflow with 25 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	26.743s (-3.1%)	27.051s (-3.6%)	0.308s	3	1.00x
🌐 Redis	Next.js (Turbopack)	26.760s	27.049s	0.289s	3	1.00x
💻 Local	Next.js (Turbopack)	27.128s	28.053s	0.925s	3	1.01x
🐘 Postgres	Express	27.240s (~)	28.066s (~)	0.826s	3	1.02x
🐘 Postgres	Nitro	27.277s (~)	28.065s (~)	0.788s	3	1.02x
💻 Local	Nitro	27.583s (~)	28.053s (~)	0.470s	3	1.03x
🌐 MongoDB	Next.js (Turbopack)	30.482s	31.045s	0.564s	2	1.14x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	46.028s (+2.9%)	47.737s (+4.3%)	1.709s	2	1.00x
▲ Vercel	Next.js (Turbopack)	50.377s (+6.6% 🔺)	52.253s (+8.0% 🔺)	1.877s	2	1.09x
▲ Vercel	Express	53.606s (+21.0% 🔺)	55.319s (+22.1% 🔺)	1.713s	2	1.16x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

workflow with 50 sequential steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	53.541s	54.093s	0.552s	2	1.00x
🐘 Postgres	Express	54.300s (~)	55.094s (~)	0.793s	2	1.01x
🐘 Postgres	Nitro	54.476s (~)	55.105s (~)	0.629s	2	1.02x
💻 Local	Express	54.851s (-3.3%)	55.101s (-3.5%)	0.250s	2	1.02x
💻 Local	Next.js (Turbopack)	55.951s	56.100s	0.149s	2	1.05x
💻 Local	Nitro	56.657s (~)	57.106s (~)	0.450s	2	1.06x
🌐 MongoDB	Next.js (Turbopack)	60.675s	61.065s	0.390s	2	1.13x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	94.771s (-4.4%)	96.975s (-4.2%)	2.204s	1	1.00x
▲ Vercel	Nitro	96.605s (~)	98.384s (~)	1.779s	1	1.02x
▲ Vercel	Next.js (Turbopack)	98.753s (-3.9%)	100.796s (-2.8%)	2.043s	1	1.04x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

Promise.all with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	1.345s	2.006s	0.662s	15	1.00x
🐘 Postgres	Express	1.366s (-4.0%)	2.011s (~)	0.645s	15	1.02x
💻 Local	Express	1.467s (-5.2% 🟢)	2.006s (~)	0.538s	15	1.09x
💻 Local	Nitro	1.498s (-1.1%)	2.005s (~)	0.507s	15	1.11x
💻 Local	Next.js (Turbopack)	1.547s	2.006s	0.459s	15	1.15x
🌐 MongoDB	Next.js (Turbopack)	2.146s	3.009s	0.863s	10	1.60x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-
🐘 Postgres	Nitro	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.312s (-4.2%)	3.407s (-4.8%)	1.095s	9	1.00x
▲ Vercel	Next.js (Turbopack)	2.641s (-1.3%)	4.038s (+7.5% 🔺)	1.397s	8	1.14x
▲ Vercel	Express	2.648s (+5.0%)	4.138s (+13.0% 🔺)	1.490s	8	1.15x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Promise.all with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	2.567s	3.008s	0.440s	10	1.00x
🐘 Postgres	Nitro	2.595s (~)	3.014s (~)	0.419s	10	1.01x
💻 Local	Express	2.600s (-14.1% 🟢)	3.007s (-15.6% 🟢)	0.407s	10	1.01x
🐘 Postgres	Express	2.615s (~)	3.015s (~)	0.399s	10	1.02x
💻 Local	Next.js (Turbopack)	3.041s	3.760s	0.718s	8	1.18x
💻 Local	Nitro	3.121s (+7.5% 🔺)	3.565s (+11.1% 🔺)	0.444s	9	1.22x
🌐 MongoDB	Next.js (Turbopack)	4.747s	5.179s	0.432s	6	1.85x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Next.js (Turbopack)	2.799s (+7.7% 🔺)	4.348s (+21.1% 🔺)	1.550s	8	1.00x
▲ Vercel	Nitro	3.520s (+30.2% 🔺)	4.708s (+23.4% 🔺)	1.188s	7	1.26x
▲ Vercel	Express	3.663s (+47.9% 🔺)	5.183s (+40.8% 🔺)	1.520s	8	1.31x

🔍 Observability: Next.js (Turbopack) | Nitro | Express

Promise.all with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	4.027s (+1.4%)	4.588s (+3.2%)	0.561s	7	1.00x
🐘 Postgres	Nitro	4.032s (+0.7%)	4.590s (+3.1%)	0.558s	7	1.00x
🌐 Redis	Next.js (Turbopack)	4.081s	5.012s	0.931s	6	1.01x
💻 Local	Express	6.787s (-15.7% 🟢)	7.014s (-20.1% 🟢)	0.227s	5	1.69x
💻 Local	Next.js (Turbopack)	7.888s	8.517s	0.629s	4	1.96x
💻 Local	Nitro	8.207s (-1.6%)	8.521s (-5.5% 🟢)	0.314s	4	2.04x
🌐 MongoDB	Next.js (Turbopack)	10.067s	10.684s	0.617s	3	2.50x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Next.js (Turbopack)	5.520s (+44.3% 🔺)	7.146s (+37.1% 🔺)	1.626s	5	1.00x
▲ Vercel	Nitro	7.009s (+153.3% 🔺)	8.352s (+119.6% 🔺)	1.343s	4	1.27x
▲ Vercel	Express	10.371s (+264.0% 🔺)	12.346s (+209.2% 🔺)	1.975s	3	1.88x

🔍 Observability: Next.js (Turbopack) | Nitro | Express

Promise.race with 10 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🌐 Redis	🥇 Next.js (Turbopack)	1.303s	2.006s	0.703s	15	1.00x
🐘 Postgres	Nitro	1.422s (-1.3%)	2.011s (~)	0.589s	15	1.09x
🐘 Postgres	Express	1.442s (-1.5%)	2.011s (-3.2%)	0.570s	15	1.11x
💻 Local	Express	1.477s (-2.9%)	2.005s (~)	0.528s	15	1.13x
💻 Local	Nitro	1.523s (-1.1%)	2.006s (~)	0.483s	15	1.17x
💻 Local	Next.js (Turbopack)	1.564s	2.073s	0.509s	15	1.20x
🌐 MongoDB	Next.js (Turbopack)	2.174s	3.009s	0.835s	10	1.67x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.131s (-0.9%)	3.290s (-1.5%)	1.159s	10	1.00x
▲ Vercel	Next.js (Turbopack)	2.268s (-27.5% 🟢)	3.823s (-24.1% 🟢)	1.555s	8	1.06x
▲ Vercel	Express	2.591s (+17.2% 🔺)	4.029s (+17.2% 🔺)	1.438s	8	1.22x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Promise.race with 25 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Express	2.526s (-3.2%)	3.012s (~)	0.486s	10	1.00x
🌐 Redis	Next.js (Turbopack)	2.532s	3.008s	0.477s	10	1.00x
🐘 Postgres	Nitro	2.730s (+7.4% 🔺)	3.015s (~)	0.285s	10	1.08x
💻 Local	Express	2.797s (-9.8% 🟢)	3.108s (-17.3% 🟢)	0.311s	10	1.11x
💻 Local	Next.js (Turbopack)	3.030s	3.760s	0.729s	8	1.20x
💻 Local	Nitro	3.124s (-1.8%)	3.886s (~)	0.762s	8	1.24x
🌐 MongoDB	Next.js (Turbopack)	4.728s	5.177s	0.449s	6	1.87x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.270s (-29.0% 🟢)	3.269s (-37.5% 🟢)	0.998s	10	1.00x
▲ Vercel	Express	2.378s (-32.1% 🟢)	3.827s (-17.2% 🟢)	1.449s	8	1.05x
▲ Vercel	Next.js (Turbopack)	3.093s (-5.5% 🟢)	4.610s (~)	1.517s	7	1.36x

🔍 Observability: Nitro | Express | Next.js (Turbopack)

Promise.race with 50 concurrent steps

💻 Local Development

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
🐘 Postgres	🥇 Nitro	3.975s (-3.5%)	4.590s (~)	0.615s	7	1.00x
🐘 Postgres	Express	4.039s (+2.3%)	4.590s (+3.2%)	0.551s	7	1.02x
🌐 Redis	Next.js (Turbopack)	4.198s	4.725s	0.527s	7	1.06x
💻 Local	Express	7.733s (-14.1% 🟢)	8.267s (-10.8% 🟢)	0.533s	4	1.95x
💻 Local	Next.js (Turbopack)	8.255s	8.766s	0.512s	4	2.08x
💻 Local	Nitro	8.519s (+0.8%)	9.275s (+2.8%)	0.756s	4	2.14x
🌐 MongoDB	Next.js (Turbopack)	10.032s	10.680s	0.648s	3	2.52x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Nitro	2.724s (-27.3% 🟢)	3.647s (-28.7% 🟢)	0.923s	9	1.00x
▲ Vercel	Next.js (Turbopack)	3.571s (-3.3%)	5.369s (+9.7% 🔺)	1.798s	6	1.31x
▲ Vercel	Express	3.789s (+1.4%)	5.100s (-0.7%)	1.311s	6	1.39x

🔍 Observability: Nitro | Next.js (Turbopack) | Express

Stream Benchmarks(includes TTFB metrics)

workflow with stream

💻 Local Development

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
💻 Local	🥇 Express	0.138s (-32.2% 🟢)	1.003s (~)	0.009s (-21.6% 🟢)	1.015s (~)	0.877s	10	1.00x
💻 Local	Next.js (Turbopack)	0.170s	1.002s	0.011s	1.017s	0.847s	10	1.24x
🌐 Redis	Next.js (Turbopack)	0.182s	1.000s	0.002s	1.007s	0.825s	10	1.33x
💻 Local	Nitro	0.212s (+7.3% 🔺)	1.003s (~)	0.011s (-2.7%)	1.017s (~)	0.804s	10	1.54x
🐘 Postgres	Express	0.215s (+2.8%)	0.996s (~)	0.002s (+15.4% 🔺)	1.013s (~)	0.797s	10	1.57x
🐘 Postgres	Nitro	0.246s (+9.4% 🔺)	0.993s (~)	0.002s (+41.7% 🔺)	1.014s (~)	0.767s	10	1.79x
🌐 MongoDB	Next.js (Turbopack)	0.474s	0.979s	0.002s	1.009s	0.534s	10	3.45x
🐘 Postgres	Next.js (Turbopack)	⚠️missing	-	-	-	-	-

▲ Production (Vercel)

World	Framework	Workflow Time	TTFB	Slurp	Wall Time	Overhead	Samples	vs Fastest
▲ Vercel	🥇 Express	1.606s (+1.9%)	2.405s (-8.7% 🟢)	0.006s (-26.2% 🟢)	3.035s (-2.2%)	1.429s	10	1.00x
▲ Vercel	Nitro	1.658s (+1.2%)	2.506s (-14.3% 🟢)	0.504s (+8896.4% 🔺)	3.490s (+1.5%)	1.833s	10	1.03x
▲ Vercel	Next.js (Turbopack)	1.785s (-4.5%)	3.025s (+6.5% 🔺)	0.011s (+160.5% 🔺)	3.640s (+6.9% 🔺)	1.855s	10	1.11x

🔍 Observability: Express | Nitro | Next.js (Turbopack)

Summary

Fastest Framework by World

Winner determined by most benchmark wins

World	🥇 Fastest Framework	Wins
💻 Local	Express	12/12
🐘 Postgres	Express	7/12
▲ Vercel	Nitro	7/12

Fastest World by Framework

Winner determined by most benchmark wins

Framework	🥇 Fastest World	Wins
Express	💻 Local	6/12
Next.js (Turbopack)	🌐 Redis	9/12
Nitro	🐘 Postgres	6/12

Column Definitions

• Workflow Time: Runtime reported by workflow (completedAt - createdAt) - primary metric
• TTFB: Time to First Byte - time from workflow start until first stream byte received (stream benchmarks only)
• Slurp: Time from first byte to complete stream consumption (stream benchmarks only)
• Wall Time: Total testbench time (trigger workflow + poll for result)
• Overhead: Testbench overhead (Wall Time - Workflow Time)
• Samples: Number of benchmark iterations run
• vs Fastest: How much slower compared to the fastest configuration for this benchmark

Worlds:

• 💻 Local: In-memory filesystem world (local development)
• 🐘 Postgres: PostgreSQL database world (local development)
• ▲ Vercel: Vercel production/preview deployment
• 🌐 Turso: Community world (local development)
• 🌐 MongoDB: Community world (local development)
• 🌐 Redis: Community world (local development)
• 🌐 Jazz: Community world (local development)

---

📋 View full workflow run

[pranaygp]

Suggested change

	'Run already finished during setup, skipping',
	'Run already finished during setup, skipping replay',

maybe this would be clearer to the user? (assuming the codepath is only for run replays)

[pranaygp]

or "skipping redundant workflow execution"

[pranaygp]

human: LGTM! great detailed state check and ty for posting that on github PR review comment