Repository navigation
fix(agent-sre): evaluate burn alerts over their declared windows - #4210
Conversation
Signed-off-by: Naveen Chatlapalli <[email protected]>
|
Azure Pipelines: There may be pipelines that require an authorized user to comment /azp run to run. |
|
🟡 Contributor Check: MEDIUM
Automated check by AGT Contributor Check. |
PR Review Summary
Verdict: AI review comments are untrusted advisory output. The summary reports workflow-generated completion status only, not model-authored pass/fail claims. |
|
Verified in code and locally: a two-hour-old error produced no firing alert on main because the burn rate was evaluated over the one-hour default; each alert now uses its declared window, so SLO status reaches WARNING or CRITICAL for one-to-24-hour-old errors, which is the intended change and worth a changelog line. Sign-off is present and CI is approved and running. Before I approve, two process items that apply to all nine PRs in this batch: please use the repository template with its Checklist, Attribution, AI Assistance and IP sections, and tick the attestations in your own words. The AI note says Codex prepared and submitted the changes at your request; CONTRIBUTING asks that no PR be submitted by an agent without a human reviewing the specific changes, so one sentence confirming you reviewed this diff before it was opened settles it. Once that and the gated checks are green I will approve and merge. |
Signed-off-by: Naveen Chatlapalli <[email protected]>
|
I reviewed the specific changes in this diff before the PR was opened. I have updated the description to use the repository template sections, completed the attestations with the validation scope and limitations, and clarified the AI-assistance note. Commit 7da0e6b adds the requested Agent-SRE changelog entry explaining the WARNING/CRITICAL status correction for errors within the declared alert window. Documentation link and frontmatter checks pass. |
MohammadHaroonAbuomar
left a comment
There was a problem hiding this comment.
Approving at 7da0e6b. Burn alerts evaluate errors over each alert's declared window instead of the default hour; new test fails on main and passes here; changelog entry added; ruff clean. Description uses the repository template with the attestations completed, and the author confirmed reviewing the diff before opening the PR. All 13 pull_request workflow runs on this head completed successfully. Commits are signed off.
Description
ErrorBudget declares 24-hour burn-rate alerts, but
firing_alerts()callsburn_rate()with its one-hour default. An error two hours ago appears in the reported 24-hour burn rate while neither alert fires.Evaluate each alert using its own
window_seconds. Thresholds and burn-rate arithmetic remain unchanged.Validation
The regression fails before the fix, covers a two-hour-old error, and confirms expiration beyond 24 hours. Objective tests: 21 passed. Full package run: 1,495 passed, 3 skipped, 1 known baseline failure (the separate Windows benchmark clock bug).
All regressions were reproduced on origin/main
406ddd4dbefore the production change. Whitespace checks pass. Ruff passes on the changed files. Changed-line spelling passes with the existing API identifierDeduplicatorallowed in the temporary check input.Scope: the existing Agent-SRE source tree, which remains exercised by repository CI. Combined validation of these six independent SRE fixes: 1,505 passed, 3 skipped on Windows/Python 3.12.
The Agent-SRE changelog now records that errors from one to 24 hours ago can correctly raise SLO status to WARNING or CRITICAL when the configured alert window includes them. Documentation link and frontmatter checks pass after this addition.
Type of Change
Package(s) Affected
Checklist
Attribution & Prior Art
Prior art / related projects: Uses the existing Agent-SRE implementation and standard library operations. No code was copied or adapted from an external project.
AI Assistance
I reviewed the specific changes in this diff before this PR was opened.
Codex assisted with implementation, regression tests, and local validation under my direction.
IP, Patents, and Licensing
Related Issues
Newly reproduced defect; the reproduction and regression coverage are described above.