monitoring-stack

name	Monitoring Stack
description	Observability infrastructure: metrics, logs, traces, alerts. Use when: setting up monitoring, creating dashboards, or configuring alerts.

Monitoring Stack

Observability infrastructure for production systems.

Three Pillars

┌─────────┐    ┌─────────┐    ┌─────────┐
│ Metrics │    │  Logs   │    │  Traces │
│  (How)  │    │ (What)  │    │ (Why)   │
└────┬────┘    └────┬────┘    └────┬────┘
     │                │                │
     ▼                ▼                ▼
  Prometheus       Loki/ELK        Jaeger
  Grafana          Grafana          Grafana

For Mansoni Platform

Metrics to Track

Category	Metrics
Frontend	FCP, LCP, CLS, TTFB, Error rate
API	Request latency p50/p95/p99, Error rate, Throughput
Database	Query latency, Connection pool, Replication lag
Infrastructure	CPU, Memory, Disk I/O, Network

Key Dashboards

1. SLO Dashboard

   • API Availability: 99.9%
   • Error Budget: 10 min/week
   • P50/P95/P99 Latency

2. User Experience

   • Core Web Vitals distribution
   • Conversion funnel drop-off
   • Feature adoption

3. Business Metrics

   • DAU/MAU
   • Revenue per user
   • Support tickets

Alerting Rules

# Example Prometheus alerting
groups:
  - name: mansoni
    rules:
      - alert: HighErrorRate
        expr: rate(http_errors_total[5m]) > 0.05
        for: 5m
        labels:
          severity: critical
        annotations:
          summary: High error rate detected

      - alert: SlowAPI
        expr: histogram_quantile(0.95, api_latency) > 2
        for: 10m
        labels:
          severity: warning

For Supabase

Monitor via Supabase Dashboard:

• Database metrics
• Auth usage
• Storage usage
• Edge function invocations
• Realtime connections

For Mansoni

Current stack:

• Sentry (errors + performance)
• Supabase Dashboard (backend metrics)
• Vercel Analytics (frontend performance)