fix: count retrying queue events separately from processing

[Divine-mercyx]

Summary

Give failed-but-retryable events their own retrying stats bucket so an outage-driven retry backlog no longer looks like healthy in-flight work.

Related Issue

Closes #205

Changes Made

• EventQueue.getStats() counts status='failed' with attempts < maxRetries as retrying instead of processing
• Heartbeat pulse now includes eventQueue stats so operators can see the retry bucket
• Queue stats type is exported; docs and health-check examples include retrying
• Test covers a mix of processing / failed-retrying / failed-exhausted events

Testing Performed

• Unit tests pass (cd engine-bridge && npm test -- event-queue.test.ts heartbeat.test.ts — 19 passed)
• Integration tests pass
• Manual smoke test performed

Checklist

• Branch follows naming convention (feat/, fix/, chore/, etc.)
• Closes #205 reference is present above (linking this PR to the issue)
• Code follows the project's style and conventions
• New or updated tests cover the changes
• Documentation updated if behaviour changed
• No secrets, credentials, or sensitive data committed
• PR is scoped to a single concern (one issue per PR)

Made with Cursor

[N-thnI]

Merged — thanks for the contribution! If you've got a moment, a rating on GrantFox helps the project.