Harden worker reconnects and runtime reporting

[mjc]

Stack 12 of 13. Depends on 10-weighted-crf-scheduling.

This preserves complete command stdout, moves worker-owned data to the system temp directory, reconnects dead Phoenix channels, normalizes every connection-loss path, discards stale reconnect attempts, reports temp-filesystem capacity, and returns typed CRF failure reasons.

Long-running workers now treat socket closure, channel errors, pull errors, and acknowledgement deadlines consistently. Capacity reporting reflects the filesystem actually used for worker data, and failed CRF searches explain why no candidate was acceptable.

Tests added in this branch cover complete stdout with bounded stderr, system-temp input placement, normalized disconnect frames, channel and pull reconnect decisions, acknowledgement liveness deadlines, stale reconnect handling, filesystem capacity payloads, and structured no-acceptable-CRF reporting including the oversized adjacent candidate.

[coderabbitai]

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: ASSERTIVE

Plan: Pro Plus

Run ID: fd25e4b8-f647-4be3-9bc7-d89fed136950

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

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.}