KAT-3696 Log the full cause when provider event ingestion fails

[gannonh]

Fixes KAT-3696.

The orchestration V2 provider event ingestion failed warning logged cause as a Cause object, which the log formatter prints as cause: { failures: [ [Object] ] }. First turns on new threads intermittently fail with "provider event stream closed unexpectedly" on the nightly, and this log was the only place the underlying error could appear. It now logs Cause.pretty(cause).

No behavior change beyond the log field. Does not fix the intermittent failure; that needs the real cause from the next occurrence.

Verified: new test logs the rendered cause when provider event ingestion fails fails on the old code (expected 'object' to equal 'string') and passes with the change; the full RunExecutionService.test.ts passes (48 tests); tsc --noEmit reports no errors.

🤖 Generated with Claude Code

[linear-code]

KAT-3696

[coderabbitai]

Important

Draft PR not reviewed

Draft PRs are not automatically reviewed by default.

• Trigger a manual review

To automatically review draft PRs, update your CodeRabbit configuration:

reviews:
  auto_review:
    drafts: true

• Autopilot · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

---

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.}