fix(metric): avoid metrics mux conflict across instances

[Ecattea]

Use per-instance ServeMux in Run and add regression test for multi-instance panic.

---

• I understand that contributing to this repository will require me to agree with the CLA

Description

This PR fixes a panic when multiple metric instances are created in the same process.

Root cause:

• metric.Run used global http.DefaultServeMux via http.Handle("/metrics", ...) and http.ListenAndServe(addr, nil).
• Re-registering /metrics across instances could panic with pattern conflict.

Changes:

• Use per-instance mux in metric.Run:
  • mux := http.NewServeMux()
  • mux.Handle("/metrics", ...)
  • http.ListenAndServe(addr, mux)
• Make startup error channel buffered to avoid potential goroutine blocking after timeout.
• Add regression test TestRun_DoesNotPanicWithMultipleInstances.

Validation:

• go test ./metric -run TestRun_DoesNotPanicWithMultipleInstances -count=50 -v
• go test ./metric -count=1 -v

What type of PR is this? (check all applicable)

• 🍕 Feature
• 🐛 Bug Fix
• 📝 Documentation Update
• 🎨 Style
• 🧑‍💻 Code Refactor
• 🔥 Performance Improvements
• ✅ Test
• 🤖 Build
• 🔁 CI

Related Tickets & Documents

• Closes Panic on registering /metrics globally #373

Mobile & Desktop Screenshots/Recordings

N/A (backend-only change)

Added tests?

• 👍 yes
• 🙅 no, because they aren't needed
• 🙋 no, because I need help

Added to documentation?

• 📜 README.md
• 📓 tech-docs
• 🙅 no documentation needed

[optional] Are there any post-deployment tasks we need to perform?

None.

[github-actions]

CLA Assistant Lite bot All contributors have signed the CLA ✍️ ✅

[Ecattea]

I have read the CLA Document and I hereby sign the CLA

[Ecattea]

The failing step check looks like a fork-permission issue, not a code/test issue.