[TRTLLMINF-37][feat] Add CI agent failure analysis to L0_MergeRequest…#12543
Open
dpitman-nvda wants to merge 2 commits intoNVIDIA:mainfrom
Open
[TRTLLMINF-37][feat] Add CI agent failure analysis to L0_MergeRequest…#12543dpitman-nvda wants to merge 2 commits intoNVIDIA:mainfrom
dpitman-nvda wants to merge 2 commits intoNVIDIA:mainfrom
Conversation
… post block Signed-off-by: Derek Pitman <dpitman@nvidia.com>
Contributor
📝 WalkthroughWalkthroughA failure post-condition handler was added to the Jenkins pipeline that invokes a failure analysis utility function and logs its results, with exception handling to prevent failure analysis errors from affecting pipeline outcome. Changes
Estimated code review effort🎯 1 (Trivial) | ⏱️ ~3 minutes 🚥 Pre-merge checks | ✅ 3✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@jenkins/L0_MergeRequest.groovy`:
- Around line 1382-1388: The call to
trtllm_utils.analyzePipelineFailureWithAgent(this, env.JOB_NAME,
env.BUILD_NUMBER) can hang; wrap that call in a Jenkins timeout block (e.g.
timeout(time: 30, unit: 'SECONDS') { ... }) and handle a timeout by catching the
TimeoutExceededException/Exception inside the existing catch so the analysis
remains best-effort and does not fail or block pipeline completion; update the
echo/log path to note a timeout occurred if triggered and keep the outer catch
to swallow any other errors.
- Around line 1384-1385: The code currently echoes the raw variable analysis
directly to the Jenkins log; redact sensitive values first by creating and using
a sanitizer function (e.g., redactSensitive or sanitizeAnalysis) that scans the
analysis string for secrets, tokens, emails, IPs, and other patterns and
replaces them with safe placeholders, then call echo "=== CI Agent Failure
Analysis ===\n${sanitizedAnalysis}" instead of echoing analysis; update the
block that references analysis to invoke this sanitizer (or a Jenkins
credential-aware redaction util) before logging and ensure tests/examples
exercise the sanitizer on the analysis variable.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Run ID: 87024dca-5ece-409e-869c-f6f13be19a56
📒 Files selected for processing (1)
jenkins/L0_MergeRequest.groovy
Comment on lines
+1382
to
+1388
| def analysis = trtllm_utils.analyzePipelineFailureWithAgent(this, env.JOB_NAME, env.BUILD_NUMBER) | ||
| if (analysis) { | ||
| echo "=== CI Agent Failure Analysis ===\n${analysis}" | ||
| } | ||
| } catch (Exception e) { | ||
| // Analysis is best-effort; do not fail the pipeline | ||
| } |
Contributor
There was a problem hiding this comment.
Add a timeout around the agent analysis call in post-failure.
On Line 1382, this external/best-effort call can hang and delay pipeline completion if the agent/dependency stalls.
Proposed fix
failure {
script {
try {
- def analysis = trtllm_utils.analyzePipelineFailureWithAgent(this, env.JOB_NAME, env.BUILD_NUMBER)
- if (analysis) {
- echo "=== CI Agent Failure Analysis ===\n${analysis}"
+ timeout(time: 2, unit: 'MINUTES') {
+ def analysis = trtllm_utils.analyzePipelineFailureWithAgent(this, env.JOB_NAME, env.BUILD_NUMBER)
+ if (analysis) {
+ echo "=== CI Agent Failure Analysis ===\n${analysis}"
+ }
}
} catch (Exception e) {
// Analysis is best-effort; do not fail the pipeline
+ echo "CI agent failure analysis skipped: ${e.toString()}"
}
}
}📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| def analysis = trtllm_utils.analyzePipelineFailureWithAgent(this, env.JOB_NAME, env.BUILD_NUMBER) | |
| if (analysis) { | |
| echo "=== CI Agent Failure Analysis ===\n${analysis}" | |
| } | |
| } catch (Exception e) { | |
| // Analysis is best-effort; do not fail the pipeline | |
| } | |
| timeout(time: 2, unit: 'MINUTES') { | |
| def analysis = trtllm_utils.analyzePipelineFailureWithAgent(this, env.JOB_NAME, env.BUILD_NUMBER) | |
| if (analysis) { | |
| echo "=== CI Agent Failure Analysis ===\n${analysis}" | |
| } | |
| } | |
| } catch (Exception e) { | |
| // Analysis is best-effort; do not fail the pipeline | |
| echo "CI agent failure analysis skipped: ${e.toString()}" | |
| } |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@jenkins/L0_MergeRequest.groovy` around lines 1382 - 1388, The call to
trtllm_utils.analyzePipelineFailureWithAgent(this, env.JOB_NAME,
env.BUILD_NUMBER) can hang; wrap that call in a Jenkins timeout block (e.g.
timeout(time: 30, unit: 'SECONDS') { ... }) and handle a timeout by catching the
TimeoutExceededException/Exception inside the existing catch so the analysis
remains best-effort and does not fail or block pipeline completion; update the
echo/log path to note a timeout occurred if triggered and keep the outer catch
to swallow any other errors.
Comment on lines
+1384
to
+1385
| echo "=== CI Agent Failure Analysis ===\n${analysis}" | ||
| } |
Contributor
There was a problem hiding this comment.
Avoid printing raw analysis content directly to Jenkins logs.
On Line 1384, analysis may include sensitive values extracted from CI logs; redact before echoing.
Proposed fix
- if (analysis) {
- echo "=== CI Agent Failure Analysis ===\n${analysis}"
+ if (analysis) {
+ def sanitizedAnalysis = analysis
+ .replaceAll(/(?i)(token|password|secret)\s*[:=]\s*\S+/, '$1=***')
+ .take(12000)
+ echo "=== CI Agent Failure Analysis ===\n${sanitizedAnalysis}"
}📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| echo "=== CI Agent Failure Analysis ===\n${analysis}" | |
| } | |
| def sanitizedAnalysis = analysis | |
| .replaceAll(/(?i)(token|password|secret)\s*[:=]\s*\S+/, '$1=***') | |
| .take(12000) | |
| echo "=== CI Agent Failure Analysis ===\n${sanitizedAnalysis}" | |
| } |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@jenkins/L0_MergeRequest.groovy` around lines 1384 - 1385, The code currently
echoes the raw variable analysis directly to the Jenkins log; redact sensitive
values first by creating and using a sanitizer function (e.g., redactSensitive
or sanitizeAnalysis) that scans the analysis string for secrets, tokens, emails,
IPs, and other patterns and replaces them with safe placeholders, then call echo
"=== CI Agent Failure Analysis ===\n${sanitizedAnalysis}" instead of echoing
analysis; update the block that references analysis to invoke this sanitizer (or
a Jenkins credential-aware redaction util) before logging and ensure
tests/examples exercise the sanitizer on the analysis variable.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
… post block
Summary by CodeRabbit
Description
In an effort to improve visibility into and automation of issues with TRT-LLM CI infrastructure, we are integrating an agentic analyzer that is triggered on failures. This agent reads through the logs of the CI run and produces a report we can review internally.
Test Coverage
N/A, this is a CI change
PR Checklist
Please review the following before submitting your PR:
PR description clearly explains what and why. If using CodeRabbit's summary, please make sure it makes sense.
PR Follows TRT-LLM CODING GUIDELINES to the best of your knowledge.
Test cases are provided for new code paths (see test instructions)
Any new dependencies have been scanned for license and vulnerabilities
CODEOWNERS updated if ownership changes
Documentation updated as needed
Update tava architecture diagram if there is a significant design change in PR.
The reviewers assigned automatically/manually are appropriate for the PR.
Please check this after reviewing the above items as appropriate for this PR.
GitHub Bot Help
To see a list of available CI bot commands, please comment
/bot help.