Visualizing System Health

Logs tell you 'What' happened; Metrics tell you 'How' your system is feeling. Dashboards are the cockpit of your architectural engine.

1. Prometheus (The Collector)

Prometheus uses a 'Pull' model. It scrapes an `/metrics` endpoint on your .NET app every 15 seconds. It record **Quantities**: - CPU / Memory usage. - Active thread count. - HTTP request throughput and failure rates. - Database connection pool health.

2. Grafana (The Artist)

Grafana connects to Prometheus and turns the data into beautiful, real-time graphs. **Architect Tip:** A 'Red' dashboard in the office (or on a shared Slack channel) is the fastest way to alert the team that the latest deployment is causing a memory leak or a performance regression.

3. AlertManager

Don't stare at dashboards. Use **AlertManager** to trigger Slack/PagerDuty notifications when a metric crosses a dangerous threshold (e.g., 'Error rate > 5% for more than 2 minutes').

4. Interview Mastery

Q: "What are the 'Golden Signals' you monitor for every service?"

Architect Answer: "Google's 4 Golden Signals: **Latency** (time to complete), **Traffic** (demand/requests per second), **Errors** (rate of failure), and **Saturation** (how 'full' the service is, e.g., memory/CPU). If these four are healthy, the service is likely healthy."