[Opt](ai-func) Improving AI function performance by linrrzqqq · Pull Request #62494 · apache/doris

linrrzqqq · 2026-04-14T14:41:05Z

Release note

Improving the performance of AI functions through batch sending, embed controls the number of (text/file) items sent in a single batch through the variable embed_max_batch_size, and the remaining functions internally maintain a conservative context window.

The current sending format is similar to:

"input": [
{"role": "system", "content": "system_prompt here"},
{"role": "user", "content": [
{"idx": 1, "text": "xxx"},
{"idx": 2, "text": "xxx"},
]
}
]

performance:

-- AI_CLASSIFYSELECTCOUNT(*) AS total_rows,
SUM(IF(res ='science', 1, 0)) AS excepte_eq_res
FROM (
SELECT AI_CLASSIFY('deepseek-chat', str, ['science', 'sport']) AS res FROM test_str
) t;
-- before +------------+----------------+
| total_rows | excepte_eq_res |
+------------+----------------+
| 100 | 100 |
+------------+----------------+1 row inset (2 min 11.579 sec)
-- now+------------+----------------+
| total_rows | excepte_eq_res |
+------------+----------------+
| 100 | 100 |
+------------+----------------+1 row inset (10.487 sec)
-- AI_FILTERSELECTCOUNT(*) AS total_rows,
SUM(IF(res =1, 1, 0)) AS zero_res_rows
FROM (
SELECT AI_FILTER('deepseek-chat', str) AS res FROM test_str
) t;
-- before+------------+---------------+
| total_rows | zero_res_rows |
+------------+---------------+
| 100 | 0 |
+------------+---------------+1 row inset (2 min 2.979 sec)
-- now+------------+---------------+
| total_rows | zero_res_rows |
+------------+---------------+
| 100 | 0 |
+------------+---------------+1 row inset (5.007 sec)
-- EMBEDselectcount(embed('qwen-embed', str)) FROM test_str;
-- before+---------------------------------+
| count(embed('qwen-embed', str)) |
+---------------------------------+
| 100 |
+---------------------------------+1 row inset (4 min 4.888 sec)
-- nowset embed_max_batch_size =10;
+---------------------------------+
| count(embed('qwen-embed', str)) |
+---------------------------------+
| 100 |
+---------------------------------+1 row inset (23.424 sec)
-- Multimodal_EmbedSELECTCOUNT(EMBED('qwen_mul_embed', to_json(file))) FROM test_jpg2;
-- before: can't get results for a long time(over 20 mins).-- nowset embed_max_batch_size =20;
+----------------------------------------------------+
| .... |
| 1152 |
+----------------------------------------------------+1142 rows inset (1 min 13.577 sec)

Check List (For Author)

Test
- Regression test
- Unit Test
- Manual test (add detailed scripts or steps below)
- No need to test or manual test. Explain why:
  - This is a refactor/code format and no logic has been changed.
  - Previous test can cover this change.
  - No code files have been changed.
  - Other reason
Behavior changed:
- No.
- Yes.
Does this need documentation?
- No.
- Yes.

Check List (For Reviewer who merge this PR)

Confirm the release note
Confirm test cases
Confirm document
Add branch pick label

hello-stephen · 2026-04-14T14:41:12Z

Thank you for your contribution to Apache Doris.
Don't know what should be done next? See How to process your PR.

Please clearly describe your PR:

What problem was fixed (it's best to include specific error reporting information). How it was fixed.
Which behaviors were modified. What was the previous behavior, what is it now, why was it modified, and what possible impacts might there be.
What features were added. Why was this function added?
Which code was refactored and why was this part of the code refactored?
Which functions were optimized and what is the difference before and after the optimization?

linrrzqqq · 2026-04-14T15:01:03Z

run buildall

hello-stephen · 2026-04-14T16:55:40Z

FE UT Coverage Report

Increment line coverage 100.00% (2/2) 🎉
Increment coverage report
Complete coverage report

hello-stephen · 2026-04-14T18:53:46Z

FE Regression Coverage Report

Increment line coverage 100.00% (2/2) 🎉
Increment coverage report
Complete coverage report

linrrzqqq · 2026-04-15T01:29:12Z

/review

github-actions

Found 1 blocking issue.

be/src/exprs/function/ai/embed.h: text embedding batching breaks Gemini resources. _execute_text_embed() now batches multiple prompts into one build_embedding_request(inputs, ...) call, but GeminiAdapter::build_embedding_request() still serializes them into a single content object and parse_embedding_response() still returns a single embedding for the text path. Running embed() on multiple rows with a GEMINI AI resource will now fail the cardinality check (expected N got 1) or effectively only embed the last input. Please either keep Gemini on the old per-row path or implement Gemini's true batch text embedding protocol before enabling batching here.

Critical checkpoint conclusions:

Goal and correctness: The PR aims to improve AI-function performance via batching. That is only partially achieved because multi-row text embed() is no longer correct for all supported providers. Existing tests do not cover the failing Gemini text-embedding path.
Scope/minimality: The change is focused, but the generic text-embedding batching applies to providers with different protocol semantics.
Concurrency: No new thread-safety or locking issue identified; the path remains synchronous.
Lifecycle/static init: No special lifecycle or static initialization issue found.
Configuration: multimodal_embed_max_batch_file_count is added and forwarded to BE correctly through TQueryOptions.
Compatibility: No storage-format or persistence compatibility issue found; FE/BE query-option propagation looks complete for the new variable.
Parallel paths: Multimodal embedding and string AI functions were updated, but the provider-specific Gemini text embedding path was not handled consistently.
Special conditions/checks: The new multimodal input validation is reasonable.
Test coverage: Unit coverage improved, but there is no test for multi-row embed() with a GEMINI resource, which is the broken path here.
Test result files: Not applicable.
Observability: Existing observability is sufficient for this review; no blocker here.
Transaction/persistence/data writes/FE-BE variable passing: Not applicable beyond query-option forwarding, which is covered.
Performance: Batching should help supported providers, but this regression must be fixed first.
Other issues: No additional blocking issue confirmed beyond the Gemini regression.

linrrzqqq · 2026-04-15T04:25:01Z

run buildall

hello-stephen · 2026-04-15T07:58:07Z

FE Regression Coverage Report

Increment line coverage 6.25% (2/32) 🎉
Increment coverage report
Complete coverage report

hello-stephen · 2026-04-15T08:24:06Z

BE UT Coverage Report

Increment line coverage 75.90% (337/444) 🎉

Increment coverage report
Complete coverage report

Category	Coverage
Function Coverage	53.08% (20174/38008)
Line Coverage	36.65% (189966/518326)
Region Coverage	32.92% (147589/448341)
Branch Coverage	34.06% (64634/189786)

hello-stephen · 2026-04-15T08:29:48Z

BE Regression && UT Coverage Report

Increment line coverage 74.02% (302/408) 🎉

Increment coverage report
Complete coverage report

Category	Coverage
Function Coverage	73.66% (27419/37222)
Line Coverage	57.31% (296134/516722)
Region Coverage	54.64% (247204/452454)
Branch Coverage	56.21% (106993/190355)

linrrzqqq · 2026-04-15T08:39:05Z

/review

github-actions

I found one blocking issue.

Critical checkpoint: Goal / correctness / tests
Conclusion: The PR goal is to batch AI function requests for better performance, and the BE unit tests cover several batch-splitting and parsing cases. However, one provider-specific compatibility path is broken, so the implementation does not fully accomplish the goal safely.
Critical checkpoint: Scope and focus
Conclusion: The change is reasonably focused on AI batching plus the needed session/thrift plumbing.
Critical checkpoint: Concurrency
Conclusion: No new concurrency or locking risks were introduced in the modified paths. The execution remains row-batch local inside function evaluation.
Critical checkpoint: Lifecycle / static initialization
Conclusion: No applicable lifecycle or static initialization issues found in the changed code.
Critical checkpoint: Configuration changes
Conclusion: embed_max_batch_size is added as a forwarded session variable and wired into TQueryOptions. FE forwarding looks complete for the new option.
Critical checkpoint: Compatibility / protocol changes
Conclusion: FE/BE protocol compatibility for embed_max_batch_size is handled, but provider API compatibility is not fully preserved: Gemini embedding now emits a batched requests payload without any corresponding endpoint normalization or compatibility handling.
Critical checkpoint: Parallel code paths
Conclusion: Text AI functions and EMBED were both updated for batching. No missing sibling path stood out beyond the Gemini provider-specific compatibility issue below.
Critical checkpoint: Special condition checks
Conclusion: Input-size validation added for multimodal requests is appropriate and clear.
Critical checkpoint: Test coverage
Conclusion: BE unit tests were expanded well for batch parsing/splitting and multimodal request shapes. But there is no test proving Gemini works end-to-end against the configured endpoint shape after this request-format change.
Critical checkpoint: Test result files
Conclusion: Not applicable; no regression .out files changed.
Critical checkpoint: Observability
Conclusion: No new observability additions were needed for this scope.
Critical checkpoint: Transaction / persistence
Conclusion: Not applicable.
Critical checkpoint: Data writes / atomicity
Conclusion: Not applicable.
Critical checkpoint: FE-BE variable passing
Conclusion: The new embed_max_batch_size variable is forwarded to BE correctly.
Critical checkpoint: Performance
Conclusion: The batching approach should significantly improve throughput, but the Gemini compatibility regression blocks safe acceptance.
Critical checkpoint: Other issues
Conclusion: No other blocking issue confirmed beyond the inline finding.

Because of the Gemini compatibility regression, I’m requesting changes.

linrrzqqq · 2026-04-15T14:30:54Z

run buildall

hello-stephen · 2026-04-15T16:05:18Z

FE UT Coverage Report

Increment line coverage 100.00% (2/2) 🎉
Increment coverage report
Complete coverage report

hello-stephen · 2026-04-15T17:36:11Z

FE Regression Coverage Report

Increment line coverage 100.00% (2/2) 🎉
Increment coverage report
Complete coverage report

hello-stephen · 2026-04-15T17:50:01Z

BE UT Coverage Report

Increment line coverage 85.81% (399/465) 🎉

Increment coverage report
Complete coverage report

Category	Coverage
Function Coverage	53.12% (20190/38007)
Line Coverage	36.69% (190245/518463)
Region Coverage	32.96% (147815/448441)
Branch Coverage	34.10% (64737/189857)

hello-stephen · 2026-04-15T17:53:32Z

BE Regression && UT Coverage Report

Increment line coverage 78.32% (336/429) 🎉

Increment coverage report
Complete coverage report

Category	Coverage
Function Coverage	71.99% (26792/37216)
Line Coverage	55.03% (284392/516825)
Region Coverage	51.99% (235258/452532)
Branch Coverage	53.47% (101805/190412)

linrrzqqq · 2026-04-16T02:00:15Z

/review

github-actions

I found two blocking regressions in the new batching work.

append_parsed_text_result() now treats any scalar text wrapped in [] as a batch array and errors out if it is not valid JSON. That breaks legitimate scalar outputs for existing functions, for example ai_mask explicitly asks the model to emit [MSKED], and ai_agg still routes arbitrary scalar text through adapter->parse_response().
Gemini embed batching now always builds a batchEmbedContents-style request body (requests array), but normalize_endpoint() only rewrites bare .../v1 / .../v1beta roots. Existing AI resources configured with the documented explicit ...:embedContent endpoint will now send the wrong payload to the old method and fail.

Critical checkpoints:

Goal: batch AI requests for better performance. The PR largely achieves that for the happy path, but the two regressions above break existing valid inputs/configurations.
Scope/focus: mostly focused, though the shared response parser changed behavior for non-batched paths too (ai_agg).
Concurrency: no new shared-state concurrency issue found in the reviewed paths.
Lifecycle/static init: no new lifecycle or SIOF issue found.
Config: embed_max_batch_size is added and forwarded to BE correctly, but I did not see FE-side range validation.
Compatibility: Gemini embed endpoint compatibility is regressed for users with explicit :embedContent resources.
Parallel paths: the shared parser change also affects non-batched callers like ai_agg.
Tests: the new BE tests cover happy-path batching well, but they miss bracketed scalar outputs and explicit-Gemini-endpoint compatibility.

HappenLee · 2026-04-16T03:15:30Z

+ // provider response -> choices[0].message.content = "[\"1\",\"0\",\"1\"]"
+ // this helper -> appends "1", "0", "1" into `results`
+ static Status append_parsed_text_result(std::string_view text,
+ std::vector<std::string>& results) {


consider maybe we can also use std::vector<string_view> to reduce str copy

有生命周期问题, 最少也得拷贝一次

hello-stephen · 2026-04-17T11:50:51Z

Cloud UT Coverage Report

Increment line coverage 🎉

Increment coverage report
Complete coverage report

Category	Coverage
Function Coverage	78.48% (1798/2291)
Line Coverage	64.14% (32291/50345)
Region Coverage	65.10% (16254/24967)
Branch Coverage	55.60% (8684/15620)

hello-stephen · 2026-04-17T15:17:54Z

BE UT Coverage Report

Increment line coverage 92.47% (479/518) 🎉

Increment coverage report
Complete coverage report

Category	Coverage
Function Coverage	53.30% (20316/38118)
Line Coverage	36.79% (191403/520224)
Region Coverage	33.11% (148882/449697)
Branch Coverage	34.18% (65048/190294)

hello-stephen · 2026-04-17T15:20:41Z

FE Regression Coverage Report

Increment line coverage 100.00% (8/8) 🎉
Increment coverage report
Complete coverage report

hello-stephen · 2026-04-17T15:23:29Z

BE Regression && UT Coverage Report

Increment line coverage 85.24% (410/481) 🎉

Increment coverage report
Complete coverage report

Category	Coverage
Function Coverage	73.85% (27565/37326)
Line Coverage	57.58% (298606/518587)
Region Coverage	54.70% (248205/453788)
Branch Coverage	56.34% (107530/190851)

hello-stephen · 2026-04-18T09:25:30Z

BE Regression && UT Coverage Report

Increment line coverage 85.24% (410/481) 🎉

Increment coverage report
Complete coverage report

Category	Coverage
Function Coverage	73.85% (27565/37326)
Line Coverage	57.58% (298606/518587)
Region Coverage	54.70% (248205/453788)
Branch Coverage	56.34% (107530/190851)

hello-stephen · 2026-04-18T09:34:47Z

BE Regression && UT Coverage Report

Increment line coverage 85.24% (410/481) 🎉

Increment coverage report
Complete coverage report

Category	Coverage
Function Coverage	73.84% (27562/37326)
Line Coverage	57.57% (298536/518587)
Region Coverage	54.72% (248311/453788)
Branch Coverage	56.33% (107513/190851)

hello-stephen · 2026-04-18T09:37:03Z

FE Regression Coverage Report

Increment line coverage 100.00% (8/8) 🎉
Increment coverage report
Complete coverage report

hello-stephen · 2026-04-18T09:43:01Z

FE Regression Coverage Report

Increment line coverage 100.00% (8/8) 🎉
Increment coverage report
Complete coverage report

linrrzqqq · 2026-04-20T03:19:06Z

/review

github-actions

I found blocking correctness issues in the new batching/parsing paths.

append_parsed_text_result() now mixes batch parsing with scalar parsing. This breaks valid ai_agg outputs that are themselves JSON array strings, and it can also let malformed bracketed output slip through for single-row batched string functions.
The new ai_context_window_size guard for scalar AI functions is only based on raw input length. build_batch_prompt() escapes quotes, backslashes, newlines, etc., so the actual JSON prompt can exceed the configured window and still be sent as one batch.
EMBED text batching has the same serialized-size accounting bug: it only sums raw input bytes before build_embedding_request() JSON-escapes the payload.

Critical checkpoints

Goal: The PR does implement batch sending and the new session variables, and the FE->BE plumbing/tests cover the happy path. However, the current code does not fully accomplish the goal safely because the response parser and window enforcement are incorrect in edge cases.
Scope/focus: The AI batching work is mostly focused, although the patch also includes an unrelated predicate_column.h fix.
Concurrency: I did not find new lock-order or thread-safety bugs in the modified scalar/embed paths. ai_agg still relies on its existing static query-context pattern; this PR does not introduce new synchronization around it.
Lifecycle/static init: No new static-initialization-order issue found. Existing ai_agg static context lifetime behavior remains.
Config changes: embed_max_batch_size and ai_context_window_size are added, validated, forwarded to BE, and regression-tested. The blocking issue is that BE enforcement is only approximate and can undercount escaped payload size.
Compatibility: The new thrift fields are optional with defaults, and the FE/BE forwarding path is updated. I did not find a rolling-upgrade serialization blocker in this patch.
Parallel code paths: Gemini/openai normalization and multimodal embedding paths were updated consistently where needed.
Special-condition checks: The new behavior depends on heuristic text-shape detection in append_parsed_text_result(), which is not safe across both batch and scalar callers.
Test coverage: Good positive coverage for batch splitting, session-variable validation, Gemini endpoint normalization, and multimodal MIME propagation. Missing negative coverage for JSON-array literal outputs in ai_agg and for escaped-input batches that exceed the serialized window.
Observability: The new warning for batch result size mismatch is useful. No additional observability blocker found.
Transaction/persistence/data writes: Not applicable for this change.
FE/BE variable passing: Updated correctly for the new session variables.
Performance: The batching work should improve performance, but the serialized-size undercount can still trigger provider-side failures under quote/backslash-heavy inputs.

HappenLee

LGTM

github-actions · 2026-04-20T09:58:27Z

PR approved by at least one committer and no changes requested.

no need

### Release note Improving the performance of AI functions through batch sending, embed controls the number of (text/file) items sent in a single batch through the variable `embed_max_batch_size`, and the remaining functions internally maintain a conservative context window. The current sending format is similar to: ```json "input": [ {"role": "system", "content": "system_prompt here"}, {"role": "user", "content": [ {"idx": 1, "text": "xxx"}, {"idx": 2, "text": "xxx"}, ] } ] ``` performance: ```sql -- AI_CLASSIFY SELECT COUNT(*) AS total_rows, SUM(IF(res = 'science', 1, 0)) AS excepte_eq_res FROM ( SELECT AI_CLASSIFY('deepseek-chat', str, ['science', 'sport']) AS res FROM test_str ) t; -- before +------------+----------------+ | total_rows | excepte_eq_res | +------------+----------------+ | 100 | 100 | +------------+----------------+ 1 row in set (2 min 11.579 sec) -- now +------------+----------------+ | total_rows | excepte_eq_res | +------------+----------------+ | 100 | 100 | +------------+----------------+ 1 row in set (10.487 sec) -- AI_FILTER SELECT COUNT(*) AS total_rows, SUM(IF(res = 1, 1, 0)) AS zero_res_rows FROM ( SELECT AI_FILTER('deepseek-chat', str) AS res FROM test_str ) t; -- before +------------+---------------+ | total_rows | zero_res_rows | +------------+---------------+ | 100 | 0 | +------------+---------------+ 1 row in set (2 min 2.979 sec) -- now +------------+---------------+ | total_rows | zero_res_rows | +------------+---------------+ | 100 | 0 | +------------+---------------+ 1 row in set (5.007 sec) -- EMBED select count(embed('qwen-embed', str)) FROM test_str; -- before +---------------------------------+ | count(embed('qwen-embed', str)) | +---------------------------------+ | 100 | +---------------------------------+ 1 row in set (4 min 4.888 sec) -- now set embed_max_batch_size = 10; +---------------------------------+ | count(embed('qwen-embed', str)) | +---------------------------------+ | 100 | +---------------------------------+ 1 row in set (23.424 sec) -- Multimodal_Embed SELECT COUNT(EMBED('qwen_mul_embed', to_json(file))) FROM test_jpg2; -- before: can't get results for a long time(over 20 mins). -- now set embed_max_batch_size = 20; +----------------------------------------------------+ | .... | | 1152 | +----------------------------------------------------+ 1142 rows in set (1 min 13.577 sec) ```

pick: #62494

linrrzqqq force-pushed the opt-ai-pr branch from cdce21e to e31b8abCompare April 14, 2026 14:56

github-actionsBot requested changes Apr 15, 2026
View reviewed changes

Comment threadbe/src/exprs/function/ai/embed.h

linrrzqqq force-pushed the opt-ai-pr branch 3 times, most recently from 831a82e to 75d0015Compare April 15, 2026 04:22

github-actionsBot requested changes Apr 15, 2026
View reviewed changes

Comment threadbe/src/exprs/function/ai/ai_adapter.h

linrrzqqq force-pushed the opt-ai-pr branch from 75d0015 to f8543b7Compare April 15, 2026 14:24

[Opt](ai-func) Improving AI function performance
2ab26a6

linrrzqqq force-pushed the opt-ai-pr branch from f8543b7 to 2ab26a6Compare April 15, 2026 14:28

github-actionsBot requested changes Apr 16, 2026
View reviewed changes

Comment threadbe/src/exprs/function/ai/ai_adapter.h
Comment threadbe/src/exprs/function/ai/ai_functions.h

HappenLee reviewed Apr 16, 2026
View reviewed changes

Comment threadbe/src/exprs/function/ai/ai_adapter.h Outdated

HappenLee reviewed Apr 16, 2026
View reviewed changes

Comment threadbe/src/exprs/function/ai/ai_functions.h Outdated

add variable ai_context_window_size
9b321e3

github-actionsBot previously requested changes Apr 20, 2026
View reviewed changes

Comment threadbe/src/exprs/function/ai/ai_adapter.h
Comment threadbe/src/exprs/function/ai/ai_functions.h
Comment threadbe/src/exprs/function/ai/embed.h

HappenLee approved these changes Apr 20, 2026
View reviewed changes

github-actionsBot added the approved Indicates a PR has been approved by one committer. label Apr 20, 2026

zclllyybb merged commit 2a8afb9 into apache:masterApr 21, 2026
30 of 31 checks passed

mymeiyi mentioned this pull request Apr 21, 2026
[fix](wal) Fix WAL replay loading 0 rows due to columns being marked as missing #62658
Merged

linrrzqqq deleted the opt-ai-pr branch April 22, 2026 07:50

HappenLee added dev/4.0.x dev/4.1.x labels Apr 24, 2026

github-actionsBot added dev/4.0.x-conflict dev/4.1.x-conflict labels Apr 24, 2026

linrrzqqq mentioned this pull request Jun 5, 2026
branch-4.1: [Opt](ai-func) Improving AI function performance #64141
Merged

yiguolei pushed a commit that referenced this pull request Jun 8, 2026
branch-4.1: [Opt](ai-func) Improving AI function performance (#64141)
cb4a8a4
pick: #62494

yiguolei added dev/4.1.2-merged and removed dev/4.1.x dev/4.1.x-conflict dev/4.0.x dev/4.0.x-conflict labels Jun 8, 2026

yiguolei mentioned this pull request Jun 14, 2026
4.1.2 Release Notes #64485
Closed

Uh oh!

Conversation

linrrzqqq commented Apr 14, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Release note

Check List (For Author)

Check List (For Reviewer who merge this PR)

Uh oh!

hello-stephen commented Apr 14, 2026

Uh oh!

linrrzqqq commented Apr 14, 2026

Uh oh!

hello-stephen commented Apr 14, 2026

FE UT Coverage Report

Uh oh!

hello-stephen commented Apr 14, 2026

FE Regression Coverage Report

Uh oh!

linrrzqqq commented Apr 15, 2026

Uh oh!

github-actionsBot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

linrrzqqq commented Apr 15, 2026

Uh oh!

hello-stephen commented Apr 15, 2026

FE Regression Coverage Report

Uh oh!

hello-stephen commented Apr 15, 2026

BE UT Coverage Report

Uh oh!

hello-stephen commented Apr 15, 2026

BE Regression && UT Coverage Report

Uh oh!

linrrzqqq commented Apr 15, 2026

Uh oh!

github-actionsBot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

linrrzqqq commented Apr 15, 2026

Uh oh!

hello-stephen commented Apr 15, 2026

FE UT Coverage Report

Uh oh!

hello-stephen commented Apr 15, 2026

FE Regression Coverage Report

Uh oh!

hello-stephen commented Apr 15, 2026

BE UT Coverage Report

Uh oh!

hello-stephen commented Apr 15, 2026

BE Regression && UT Coverage Report

Uh oh!

linrrzqqq commented Apr 16, 2026

Uh oh!

github-actionsBot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

HappenLeeApr 16, 2026

Choose a reason for hiding this comment

Uh oh!

linrrzqqqApr 16, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

hello-stephen commented Apr 17, 2026

Cloud UT Coverage Report

Uh oh!

hello-stephen commented Apr 17, 2026

BE UT Coverage Report

Uh oh!

hello-stephen commented Apr 17, 2026

FE Regression Coverage Report

Uh oh!

linrrzqqq commented Apr 14, 2026•
edited
Loading

linrrzqqqApr 16, 2026•
edited
Loading