Repository navigation
Conversation
The image build verifies ~120 protocol archives against the CDN, each with a 60 s timeout and no retry, so one slow response fails the whole deploy (run 44353722716: "The operation was aborted due to timeout" at 64 s). Retry discovery and each archive up to twice (2 s, then 10 s) on a thrown fetch, an aborted or interrupted body, or a 5xx/429. A definitive answer, such as a 4xx, a Fly fallback, a missing R2 ETag or different bytes, still fails at once, and nothing is pruned until every body verifies. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Contributor
There was a problem hiding this comment.
Ladon verdict: Approve
Approve — bounded retry hardening on CDN verification in a build-time script.
Checked:
- Only two files touched:
scripts/prune-runtime-artifacts.cjsand its test. Nostatic/schemas/source/**,docs/reference/**,dist/**, orpackage.jsonversion changes — no changeset owed, no schema↔docs coherence concern, no oneOf discriminator impact, no released-artifact mutation. - Retry logic (3 attempts, per-attempt 60s timeout) correctly fail-closes on definitive negatives (OffloadRejected, Fly fallback, bad ETag, byte mismatch) and only retries transient failures. Timeout rebuilt per attempt so it resets; file read/hash baseline stays outside the wrapper.
- Tests cover retry-then-succeed, no-retry-on-definitive, and exhaustion.
- No Critical/High/Medium findings. No high-risk paths, no gated paths, no no-auto-approve team match, no deletions.
Decision table: rows 1–8 do not fire (no blocking/medium findings, gated_paths false, high_risk false, no team gate). Note review_decision is REVIEW_REQUIRED, but gated_paths is false so row 2 does not apply. Falls through to row 9 → approve.
This branch has not been deployed
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
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.
Why
Production deploy run 44353722716 (head
859ea919be) failed during the image build, not the rollout:Before it removes any local protocol archive,
prune-runtime-artifacts.cjs --offload-protocolfetches CDN discovery and then each archive, about 120 of them, six at a time. Every request has a 60 sAbortSignal.timeoutand no retry, so one slow response from the Fly builder fails the deploy. The code hasn't changed since #8814, and discovery answers in about 2.5 s now, so this was transient.What
withFetchRetries: up to 3 attempts, waiting 2 s and then 10 s. Each attempt keeps its own 60 s timeout. Every retry logs a warning line, so the build log shows it.OffloadRejected): any other 4xx, a Fly-fallback response (fly-request-id), an archive without an R2 strong ETag, or bytes that differ from the committed ones. Those are answers, not stalls.Tests
npm run test:runtime-artifacts: 14/14 pass. This needs the3.3.0-rc.0docs snapshot checked out.🤖 Generated with Claude Code