Monitoring and Observability: Building Confidence in Production

1. Metrics — Quantitative measurements over time (CPU usage, request latency, error rates).
2. Logs — Discrete, timestamped records of events (application errors, access records, deployment events).
3. Traces — End-to-end records of a request as it flows through distributed services.

Together, they give you a complete picture of system health and behavior.

Setting Up Prometheus and Grafana

Prometheus is a time-series metrics database that pulls (scrapes) metrics from your services. Grafana provides beautiful dashboards to visualize that data.

# prometheus.yml
global:
  scrape_interval: 15s
  evaluation_interval: 15s

scrape_configs:
  - job_name: 'prometheus'
    static_configs:
      - targets: ['localhost:9090']

  - job_name: 'node-exporter'
    static_configs:
      - targets: ['node-exporter:9100']

  - job_name: 'application'
    metrics_path: '/metrics'
    static_configs:
      - targets: ['app:8080']

Instrument your application to expose a /metrics endpoint. For a Node.js app using the prom-client library:

const client = require('prom-client');
const collectDefaultMetrics = client.collectDefaultMetrics;

collectDefaultMetrics({ prefix: 'myapp_' });

app.get('/metrics', async (req, res) => {
  res.set('Content-Type', client.register.contentType);
  res.end(await client.register.metrics());
});

Key Metrics to Track

Not all metrics are created equal. Focus on what matters:

• RED method (for services): Rate of requests, Errors, and Duration (latency).
• USE method (for resources): Utilization, Saturation, and Errors.
• Business metrics: Conversion rate, revenue per minute, active users.

A practical Grafana dashboard might include:

Structured Logging with JSON

Unstructured logs are hard to search and analyze. Switch to structured JSON logging:

{
  "timestamp": "2026-05-21T10:30:00Z",
  "level": "ERROR",
  "service": "payment-api",
  "trace_id": "abc-123-def",
  "message": "Payment processing failed",
  "error": "Gateway timeout",
  "user_id": "usr_456"
}

Tools like the ELK stack (Elasticsearch, Logstash, Kibana) or Grafana Loki excel at ingesting and querying structured logs at scale.

Setting Up Alerts

Alerts should be actionable, not noisy. Follow the rule: every alert must have a clear runbook.

# Example Prometheus alerting rule
groups:
  - name: application
    rules:
      - alert: HighErrorRate
        expr: rate(http_requests_total{status=~"5.."}[5m]) / rate(http_requests_total[5m]) > 0.05
        for: 2m
        labels:
          severity: critical
        annotations:
          summary: "High error rate detected"
          description: "Error rate is above 5% for the last 2 minutes."
          runbook: "https://wiki.internal/runbooks/high-error-rate"

Conclusion

Observability isn't a one-time setup — it's an ongoing practice. Start with the basics: collect metrics, centralize logs, and set up sensible alerts. Then iterate based on what incidents reveal about your blind spots. The goal isn't to prevent every outage; it's to reduce the time between "something is wrong" and "we know what's wrong and how to fix it."