fix(retry): context_overflow falls through to the next candidate instead of aborting

[jmlago]

Problem (diagnosed against prod v-3a2e0ea)

retry_policies.balanced.context_overflow = { action = "abort" } was the only non-stream error kind that aborts the whole request. That predates provider-neutral families (#26): a family like gpt-5.4 now spans candidates with heterogeneous context windows (openai, openrouter, antseed), so an overflow on the first route says nothing about the rest — yet abort killed the request without trying them.

Prod trace of the exact failure — family:gpt-5.4 tried only openrouter/gpt-5.4, got a 400, and aborted:

router: context_overflow — openrouter/gpt-5.4=context_overflow(400) "..."

Change

context_overflow now falls through like bad_request/timeout. retry_same would be futile (same model, same window); next_candidate is not — another route may have a larger window. If every candidate overflows the request still ends cleanly in exhausted: context_overflow.

stream_interrupted stays abort (content already delivered — can't retry).

Tests (fail against HEAD)

Behavioural, against the real config.live.lua: a family:gpt-5.4 whose every route returns context_overflow now calls more than one candidate before exhausting, where HEAD aborted on the first (reproducing the prod shape: only openrouter was tried). 45 passed across live-wiring + flow + policy-ir files.

Sibling PR: fix/legible-provider-400s (precise 400 classification + real provider message). With both, a misclassified 400 is no longer fatal and genuine overflow gets a fallback.

[coderabbitai]

Warning

Review limit reached

@jmlago, we couldn't start this review because you've reached your PR review rate limit.

More reviews will be available in 59 minutes and 48 seconds. Learn how PR review limits work.

Your organization has used up its prepaid credits, and credit purchases are no longer available. Enable the review add-on in the billing tab to keep reviews running — you're only billed for reviews past your plan's rate limits ($0.25/file).

⌛ How to resolve this issue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based credits.

🚦 How do rate limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please see our Fair Usage Limits Policy for further information.

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: a03ee41c-ff53-4b17-8beb-8bb6095a6256

📥 Commits

Reviewing files that changed from the base of the PR and between 3a2e0ea and a88febb.

📒 Files selected for processing (2)

• config.live.lua
• tests/test_live_wiring.py

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/context-overflow-falls-through

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}