Repository navigation
Add bounded repository-verified runtime mirror failover - #151
Conversation
|
@codex review |
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: e24be5e477
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
Implemented both review findings and integrated the resource-admission/checksum/offline baseline. Body failures count once and only complete EOF clears failures; stale callbacks cannot alter a newer routing epoch. Candidate replay response headers share the selection deadline, while accepted response bodies retain their ordinary request lifetime. Full race tests, vet and lint pass; coverage is 83.3%. @codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 8e183f4c01
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
Codex Review: Didn't find any major issues. 🚀 Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
|
Fixed recovered-mirror health for successful bodyless replays without clearing body-bearing failures early. Full and repeated focused race tests, vet and lint pass. Security Scan now also runs on PRs targeting feature branches, matching the existing CI trigger and keeping stacked PRs covered. @codex review |
|
Codex Review: Didn't find any major issues. Delightful! Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
An opt-in mirror-health controller observes complete public Release/InRelease responses and switches away from repeated transport, incomplete-body or 5xx failures only after a candidate matches the learned SHA256 snapshot. Candidate probes and response headers use a shared deadline, resource admission, byte limits and cooldown; accepted response bodies retain their normal request lifetime. Unknown, expired or inconsistent suites stay pinned, and pool requests require a consistent view of all observed suites.
Explicit repository IDs, approved public query parameters and a conservative header policy prevent credential or tenant data from crossing mirrors. Routing epochs and active-source checks reject stale probe/body callbacks. Add authenticated read-only
/api/mirrors/status, configuration validation and operational documentation.Depends on #154. Validation: integrated full short race suite, vet and lint pass (83.3% coverage). Focused race regressions cover concurrent selection, refresh during probes, truncated fixed/chunked responses, body read timeouts, bounded candidate headers and accepted bodies surviving the probe window.