feat(async-jobs): async execution with job queue backends

[waleedlatif1]

Summary

• Add async job queue system with Redis, database, and trigger.dev backends
• Jobs auto-cleanup via Redis TTL (24h) and cron retention cleanup
• Optimized DB indexes for cleanup queries
• Add createMockRedis() to @sim/testing

Type of Change

• New feature

Testing

• Unit tests for Redis backend and shared constants (14 tests)
• Tested manually

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing
• No new warnings introduced
• I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

^{Cursor Bugbot found 2 potential issues for commit fe401f2}

[vercel]

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

1 Skipped Deployment

Project	Deployment	Actions	Updated (UTC)
docs	Skipped		Feb 4, 2026 10:42pm

[greptile-apps]

Greptile Overview

Greptile Summary

This PR implements a robust async job queue system with three backend options (Trigger.dev, Redis, Database) using an automatic fallback chain. The implementation follows a clean adapter pattern with a unified JobQueueBackend interface.

Key Changes:

• Added async job queue infrastructure with Redis, Database, and Trigger.dev backends
• Redis backend uses TTL-based auto-cleanup (48h max lifetime, 24h retention for completed/failed)
• Database backend properly increments attempts using SQL expressions
• Unified job status API (/api/jobs/[jobId]) with access control
• Extended cron cleanup to handle stale async jobs and enforce retention policy
• Added dedicated async timeout environment variables (90 min default vs 5-50 min sync)
• Workflow execution routes now support async mode via X-Execution-Mode: async header
• Fire-and-forget execution for Redis/DB backends; Trigger.dev handles its own workers
• Comprehensive unit tests (14 tests) and new createMockRedis() test utility

Architecture:
The system uses a singleton pattern for backend selection with fallback logic: Trigger.dev → Redis → Database. For Redis/DB backends, jobs execute inline (fire-and-forget) after being queued. The cleanup cron handles both stale job detection and retention enforcement.

Database Changes:
New async_jobs table with composite indexes on (status, started_at) and (status, completed_at) optimized for cleanup queries.

Confidence Score: 4.5/5

• This PR is safe to merge with high confidence - well-architected async job system with proper cleanup
• Score reflects solid architecture, comprehensive tests, and proper retry handling. Previous concerns about Redis attempts counter were addressed (uses hincrby). Minor deduction for complexity of inline execution pattern and potential race conditions.
• Pay attention to apps/sim/app/api/workflows/[id]/execute/route.ts (complex async integration) and the inline fire-and-forget execution pattern in Redis/DB backends

Important Files Changed

Filename	Overview
apps/sim/lib/core/async-jobs/config.ts	Implements backend selection logic with fallback chain: trigger.dev → redis → database
apps/sim/lib/core/async-jobs/backends/redis.ts	Redis backend implementation with TTL-based cleanup and job state management
apps/sim/lib/core/async-jobs/backends/database.ts	Database backend using Drizzle ORM, properly increments attempts using SQL expression
packages/db/schema.ts	Added async_jobs table with composite indexes for efficient cleanup queries
apps/sim/app/api/cron/cleanup-stale-executions/route.ts	Extended cron to cleanup stale async jobs and enforce retention policy
apps/sim/app/api/workflows/[id]/execute/route.ts	Integrated async execution mode with job queue and inline fire-and-forget execution

Sequence Diagram

sequenceDiagram
participant Client
participant API as API Route<br/>/api/workflows/[id]/execute
participant Config as Job Queue Config
participant Backend as Job Backend<br/>(Redis/DB/Trigger.dev)
participant Worker as Background Worker<br/>(executeWorkflowJob)
participant Executor as Workflow Executor
Client->>API: POST with X-Execution-Mode: async
API->>Config: getJobQueue()
Config->>Config: Determine backend type<br/>(trigger.dev → redis → database)
Config-->>API: Return backend instance
API->>Backend: enqueue('workflow-execution', payload)
Backend->>Backend: Generate jobId (run_xxxx)
Backend->>Backend: Store job with status=pending
alt Redis Backend
Backend->>Backend: Set TTL=48h (max lifetime)
end
Backend-->>API: Return jobId
API->>Client: 202 Accepted<br/>{jobId, statusUrl, async: true}
alt shouldExecuteInline() == true (Redis/DB)
API->>Worker: Fire-and-forget execution
Worker->>Backend: startJob(jobId)
Backend->>Backend: Set status=processing<br/>Increment attempts
Worker->>Executor: executeWorkflowCore()
Executor-->>Worker: Execution result
alt Success
Worker->>Backend: completeJob(jobId, output)
Backend->>Backend: Set status=completed<br/>Store output
alt Redis Backend
Backend->>Backend: Set TTL=24h (retention)
end
else Failure
Worker->>Backend: markJobFailed(jobId, error)
Backend->>Backend: Set status=failed<br/>Store error
alt Redis Backend
Backend->>Backend: Set TTL=24h (retention)
end
end
else Trigger.dev Backend
Note over Backend: Trigger.dev handles execution<br/>via its own worker system
end
Client->>API: GET /api/jobs/{jobId}
API->>Backend: getJob(jobId)
Backend-->>API: Job with status & output
API-->>Client: Job status response

Loading

[greptile-apps]

_{9 files reviewed, 1 comment}

_{Edit Code Review Agent Settings | Greptile}

[waleedlatif1]

@cursor review

[waleedlatif1]

@greptile

[greptile-apps]

_{6 files reviewed, 1 comment}

_{Edit Code Review Agent Settings | Greptile}

[waleedlatif1]

@cursor review

[waleedlatif1]

@greptile

[greptile-apps]

_{8 files reviewed, no comments}

_{Edit Code Review Agent Settings | Greptile}

[waleedlatif1]

@cursor review

[waleedlatif1]

@greptile

[waleedlatif1]

@cursor review

[greptile-apps]

_{6 files reviewed, no comments}

_{Edit Code Review Agent Settings | Greptile}

[cursor]

Cursor Bugbot has reviewed your changes and found 2 potential issues.

^{Bugbot Autofix is OFF. To automatically fix reported issues with Cloud Agents, enable Autofix in the Cursor dashboard.}