feat(server): add graceful shutdown on SIGTERM/SIGINT

[tlgimenes]

What is this contribution about?

Adds graceful shutdown handling when the MCP Mesh server receives SIGTERM (K8s pod termination) or SIGINT (Ctrl+C). Previously, all resources — database connections, NATS subscriptions, event bus workers, in-flight telemetry — were abandoned without cleanup.

The shutdown sequence follows a strict order: stop HTTP servers → stop workers in parallel (EventBus, SSE hub, cron, decopilot) → drain NATS (after all consumers stopped) → flush telemetry → close database. A 10-second force-exit timeout prevents the process from hanging indefinitely.

Two-file change using Object.assign(app, { shutdown }) to preserve backward compatibility with existing tests.

How to Test

1. bun run dev, then press Ctrl+C
2. Should see [shutdown] Received SIGINT... → [shutdown] Stopping workers... → [shutdown] Cleanup complete.
3. bun run dev, then in another terminal kill -TERM <pid> — same clean shutdown behavior
4. bun run check passes, bun test apps/mesh/src/api/ passes (317 tests)

Review Checklist

• PR title is clear and descriptive
• Changes are tested and working
• No breaking changes

---

Summary by cubic

Add graceful shutdown on SIGTERM/SIGINT and separate liveness/readiness probes. Ensures clean shutdown, prevents lost data, and improves K8s draining during rollouts.

• New Features
  • Shutdown sequence: stop HTTP server first (force-close SSE), stop workers in parallel (EventBus, SSE hub, cron, decopilot), drain NATS, flush telemetry, then close the DB.
  • Added /health/live and /health/ready; readiness checks DB and NATS, returns 503 during shutdown via app.markShuttingDown(). /health kept for compatibility.
  • Installed SIGTERM/SIGINT handlers with a 2s pre-stop delay and a 55s force-exit timeout; exposes app.shutdown().
  • NATS_URL now accepts comma-separated URLs for cluster failover.
  • Helm chart bumped to 0.1.41 with new probe paths and terminationGracePeriodSeconds=60; added deploy/docker-compose/docker-compose.dev.yml for local Postgres and a 3-node NATS cluster.

^{Written for commit fee6fc4. Summary will update on new commits.}

[github-actions]

🧪 Benchmark

Should we run the Virtual MCP strategy benchmark for this PR?

React with 👍 to run the benchmark.

Reaction	Action
👍	Run quick benchmark (10 & 128 tools)

Benchmark will run on the next push after you react.

[github-actions]

Release Options

Suggested: Minor (2.203.0) — based on feat: prefix

React with an emoji to override the release type:

Reaction	Type	Next Version
👍	Prerelease	2.202.5-alpha.1
🎉	Patch	2.202.5
❤️	Minor	2.203.0
🚀	Major	3.0.0

Current version: 2.202.4

Note: If multiple reactions exist, the smallest bump wins. If no reactions, the suggested bump is used (default: minor).

[cubic-dev-ai]

1 issue found across 2 files

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="apps/mesh/src/api/app.ts">

<violation number="1" location="apps/mesh/src/api/app.ts:1156">
P2: Wrap cleanup callbacks in a deferred promise (`Promise.resolve().then(...)`) so synchronous throws are captured by `Promise.allSettled` instead of aborting shutdown early.</violation>
</file>

_{Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review.}

[tlgimenes]

merged by #2855