feat: Micrometer integration for Prometheus/Grafana metrics

[rlaope]

Problem

Argus metrics are only available through its own dashboard and API:

• Cannot integrate with existing monitoring infrastructure
• No support for standard metrics formats (Prometheus, OpenMetrics)
• Cannot create custom Grafana dashboards
• No long-term metrics storage

Proposed Solution

Integrate Micrometer for standard metrics exposure:

1. Expose metrics in Prometheus format
2. Provide pre-built Grafana dashboard
3. Support custom metric tags
4. Compatible with existing monitoring stacks

Acceptance Criteria

• Add Micrometer dependency
• Expose /prometheus endpoint with metrics
• Include all current metrics (events, threads, pinning)
• Add JVM metrics (memory, GC, threads)
• Support custom tags via configuration
• Provide Grafana dashboard JSON
• Document Prometheus scrape configuration

Metrics to Expose

# HELP argus_virtual_thread_events_total Total virtual thread events
# TYPE argus_virtual_thread_events_total counter
argus_virtual_thread_events_total{type="START"} 15234
argus_virtual_thread_events_total{type="END"} 15100
argus_virtual_thread_events_total{type="PINNED"} 45
argus_virtual_thread_events_total{type="SUBMIT_FAILED"} 0

# HELP argus_virtual_threads_active Current active virtual threads
# TYPE argus_virtual_threads_active gauge
argus_virtual_threads_active 134

# HELP argus_pinned_events_rate Pinning events per second
# TYPE argus_pinned_events_rate gauge
argus_pinned_events_rate 0.5

# HELP argus_carrier_threads_active Active carrier threads
# TYPE argus_carrier_threads_active gauge
argus_carrier_threads_active 8

# HELP argus_websocket_clients Connected WebSocket clients
# TYPE argus_websocket_clients gauge
argus_websocket_clients 2

Prometheus Configuration

# prometheus.yml
scrape_configs:
  - job_name: 'argus'
    static_configs:
      - targets: ['localhost:8080']
    metrics_path: '/prometheus'
    scrape_interval: 5s

Grafana Dashboard

Provide JSON dashboard with:

• Events rate panel (START, END, PINNED over time)
• Active threads gauge
• Pinning alerts panel
• Carrier utilization heatmap
• Top pinning hotspots table

Configuration

argus.metrics.enabled=true
argus.metrics.prometheus.enabled=true
argus.metrics.tags.application=my-app
argus.metrics.tags.environment=production

Technical Considerations

• Use micrometer-registry-prometheus
• Consider push gateway support for short-lived apps
• Add histogram for thread duration distribution
• Support dimensional metrics with thread name prefix patterns