Monitoring Setup

Set up observability with logs, metrics, and traces — know what your system is doing at all times.

RahulRachhoya Updated

File contents

Monitoring Setup

Three Pillars

  1. Logs — structured, searchable records of events
  2. Metrics — numerical measurements over time (latency, error rate, throughput)
  3. Traces — end-to-end request flow across services

What to Monitor

  • RED metrics: Rate, Errors, Duration (for every service)
  • USE metrics: Utilization, Saturation, Errors (for every resource)
  • Business metrics: signups, orders, revenue — know if the business is working

Alert Design

  • Alert on symptoms (high error rate), not causes (high CPU)
  • Set proper thresholds — too many alerts = ignored alerts
  • Page for: user-facing errors, data loss, security incidents
  • Email/Slack for: warnings, approaching thresholds
  • Every alert must have a runbook link

Logging Rules

  • Structured format (JSON), not plain text
  • Include: timestamp, level, service, trace_id, message
  • No secrets in logs (PII, passwords, tokens)
  • Log at the right level: DEBUG (dev), INFO (normal), WARN (concern), ERROR (broken)

RahulRachhoya/auto-skill-mcp/tree/main/src/auto_skill_mcp/skill_data/monitoring-setup commit f6d873c8d8

Frequently asked questions

npx skillmds@latest add rahulrachhoya/monitoring-setup