Add Prometheus metrics and health endpoints

[dorser]

Summary

Micromize has no observability features — no metrics endpoint, no health checks, no readiness/liveness probes.

What's Needed

1. Prometheus Metrics

Expose metrics at /metrics:

• micromize_events_total{gadget, action} — Counter of events by gadget and action (blocked/audited)
• micromize_gadgets_running — Gauge of currently running gadgets
• micromize_gadget_errors_total{gadget} — Counter of gadget errors
• micromize_uptime_seconds — Uptime gauge

2. Health Endpoint

Expose /healthz and /readyz endpoints:

• Healthy when all gadgets are loaded and running
• Unhealthy if a gadget fails

3. Kubernetes Probes

Add to the DaemonSet template:

livenessProbe:
  httpGet:
    path: /healthz
    port: 8080
readinessProbe:
  httpGet:
    path: /readyz
    port: 8080

Considerations

• The tool runs as a DaemonSet — metrics should be scrapable per-node
• Consider whether to use the Prometheus Go client library or a lightweight alternative
• Port should be configurable

Files

• internal/runtime/manager.go
• internal/gadget/registry.go
• charts/micromize/templates/daemonset.yaml