Skip to content

Complete Phase 4 health-gated service discovery#337

Merged
kacy merged 2 commits intomainfrom
phase4-complete
Mar 27, 2026
Merged

Complete Phase 4 health-gated service discovery#337
kacy merged 2 commits intomainfrom
phase4-complete

Conversation

@kacy
Copy link
Copy Markdown
Owner

@kacy kacy commented Mar 27, 2026

Summary

  • gate runtime endpoint eligibility on generation-aware health readiness instead of durable admin-state churn
  • replace the fixed, sequential health checker with a bounded worker-based scheduler that rejects stale probe completions
  • expose health-checker capacity, queue, and stale-result telemetry in service rollout status, metrics, and service endpoint detail

Commits

  • fabc97d Gate endpoint eligibility by health generation
  • 5155a7c Add bounded generation-aware health checking

Verification

  • zig fmt on touched files
  • bounded zig build test runs for:
    • state machine — healthy to unhealthy after retries
    • route handles /v1/status?mode=service_rollout GET
    • handleMetricsPrometheus exposes service rollout metrics
  • each timed out after 45 seconds with no emitted failures

Notes

  • probe transitions no longer rewrite durable endpoint admin state
  • readiness-gated endpoints stay ineligible until a matching-generation probe succeeds
  • stale probe completions are dropped by generation and registration epoch before they can republish service membership

@kacy kacy merged commit a0c9c3e into main Mar 27, 2026
5 of 7 checks passed
@kacy kacy deleted the phase4-complete branch March 27, 2026 12:28
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