Skip to content

fix: retain captures during worker restarts - #10

Merged
davidrobertson merged 1 commit into
mainfrom
fix/preserve-capture-during-worker-restarts
Oct 10, 2026
Merged

davidrobertson merged 1 commit into
mainfrom
fix/preserve-capture-during-worker-restarts

Conversation

@davidrobertson

Copy link
Copy Markdown
Owner

Capture hooks could lose observations during a daemon restart: the engine's HTTP listener stayed up while the worker route returned 404, and clients treated that as a permanent rejection. Hooks now spool those observations, and CLI drains retain existing records until the route returns. Validation and authentication rejections keep their existing behavior.

This follow-up publishes fork version 0.9.30-codex.2 with matching plugin manifests and export compatibility.

Validation: both new regressions failed before the fix. Final suite: 2,683 passed, zero failures, four opt-in skips. All ten packed-artifact scenarios passed. A real iii 0.22.1 engine without a worker returns404, the hook spools, the CLI keeps the record, and worker startup recovers it into searchable memory. Independent review and QA reports are bound to the final source SHA before handoff.

@davidrobertson
davidrobertson merged commit 5e9dfd4 into main Oct 10, 2026
10 checks passed
@davidrobertson
davidrobertson deleted the fix/preserve-capture-during-worker-restarts branch October 10, 2026 16:53
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant