[aw] CI Review hit engine rate limit (HTTP 429)

[github-actions]

Workflow Failure

Workflow: CI Review
Branch: main
Run: https://github.com/pydantic/pydantic-ai/actions/runs/34301874824
Pull Request: #5123

Warning

Engine Rate Limited (HTTP 429): The claude engine hit provider rate limits and could not complete this run.

This signal was detected from engine runtime logs/OTLP telemetry.

What to do next

• Retry the workflow after a short delay.
• If this keeps happening, reduce concurrent workflow volume or review provider quota/rate-limit policies.

Action Required

Assign this issue to an agent to debug and fix the issue.

Debug with any coding agent

Use this prompt with any coding agent (GitHub Copilot, Claude, Gemini, etc.):

Debug the agentic workflow failure using https://raw.githubusercontent.com/github/gh-aw/main/debug.md

The failed workflow run is at https://github.com/pydantic/pydantic-ai/actions/runs/34301874824

Manually invoke the agent

Debug this workflow failure using your favorite Agent CLI and the agentic-workflows prompt.

• Start your agent
• Load the agentic-workflows skill from .github/skills/agentic-workflows/SKILL.md or https://github.com/github/gh-aw/blob/main/.github/skills/agentic-workflows/SKILL.md
• Type debug the agentic workflow pydantic-ai-pr-review failure in https://github.com/pydantic/pydantic-ai/actions/runs/34301874824

Tip

Stop reporting this workflow as a failure

To stop a workflow from creating failure issues, set report-failure-as-issue: false in its frontmatter:

safe-outputs:
  report-failure-as-issue: false

Generated from CI Review · ◷

• expires on Sep 16, 2026, 2:31 AM UTC

[pydanty]

• Relationship resolved
• Existing-work coverage resolved
• Investigation inputs sufficient
• Supported production reachability established
• Contract and historical intent reconciled (not established)
• User/project relevance established (not applicable)
• Immediate relief assessed (not applicable)
• Long-term change is net-positive (not applicable)
• Change boundary and invasiveness understood (not applicable)
• Success criteria and integration-level regression tests defined (not applicable)

Decision: maintainer discussion

The report is fully authenticated — run 34301874824 failed with 14 verified, unretried MiniMax 429s under GH_AW_HARNESS_MAX_RETRIES: "0" — but the "should it not have happened" question is unsettled: retry=0 is a deliberate checked-in per-agent choice with an inline rationale, while PR #6853 left the #6766 engine-429 retry/backoff finding "still assigned" and no maintainer has endorsed the recurring failure issues. Contract- and history-ambiguous, third open instance of the class (#8154/#8193), and the mitigation direction (retry/backoff, report-failure-as-issue: false, class consolidation) is a maintainer policy decision, so the disposition is maintainer discussion.

Why it matters: The CI Review workflow that reviews PRs on main hard-fails whenever the MiniMax gateway rate-limits, mints a failure issue that self-expires (~09-12/#8154, ~09-14/#8193, 09-16/#8231), and no fix or consolidation exists — so the review automation is silently degraded and the same report recurs roughly every other day. A maintainer decision is needed before any code change: whether to add 429 retry/backoff per the deferred #6766 finding, opt the workflow out of failure-issue minting, consolidate the three open instances around #8154, or accept the current tradeoff; acting unilaterally risks overriding the documented stateless-shim rationale for retry=0. Secondary: the gh-aw template's hardcoded PR link (#5123, refuted across the class — run 34301874824 reviewed #8059) misroutes any future dedupe keyed on it and should be flagged to whoever owns the template.

Evidence: Issue (#8231): bot-generated gh-aw failure report for CI Review run 34301874824, category engine_rate_limit_429; its hardcoded "Pull Request: #5123" link is refuted (the run reviewed #8059) and the untrusted github/gh-aw debug/SKILL links were not fetched — trusted signal is the gh-aw footer plus the run itself. Reproduction: the run concluded failure with exactly 14× HTTP 429 from the MiniMax-routed engine gateway (ANTHROPIC_BASE_URL=api.minimax.io/anthropic), unretried because engine-minimax.md sets GH_AW_HARNESS_MAX_RETRIES: "0" for the main review agent, ending in exit code 1; the issue was minted 4 seconds after the run's final update. Contract: retry=0 is a deliberate per-agent checked-in choice with an inline stateless-shim rationale (threat-detection in the same file keeps "3"); failure-issue minting is default gh-aw tool behavior with a documented opt-out not set in the repo; no doc, test, or maintainer statement says the current behavior is wrong now. Historical intent: PR #6853 (dmontagu) consciously deferred the #6766 engine-429 retry/backoff finding as "still assigned", zero human comments exist across the ~11-issue CI Review class, and the #6766 primary text plus the provenance of retry=0 could not be verified this run — both directions remain live. Relationship: #8231 is the third open instance of one unresolved failure class (#8154 oldest, #8193), a partial duplicate whose fix would resolve the class; consolidation is itself a policy call. Existing work: none — #8201/#8205 cover gateway 401/403 auth, a different mechanism, and #6853's mitigation was deferred, not implemented. Actionability: inputs are sufficient for a maintainer decision brief covering mitigation direction, canonical consolidation, and the template's stale PR link.

Playground: https://play.pydantic.work/p/EnBJQCgA