Redis Observability

A practical reference for keeping Redis healthy in production. It lists the six metrics you should actually monitor (memory, connections, hit ratio, ops/sec, rejected connections, persistence lag) and the built-in commands that matter during an incident: SLOWLOG for finding slow queries, MEMORY DOCTOR for diagnosing pressure, CLIENT LIST for connection storms, and FT.PROFILE for search performance. The guidance is opinionated about alert thresholds (80% memory, sub-80% hit ratio) and includes minimal Python snippets for extracting what you need from INFO. Most useful when you're setting up monitoring for the first time or triaging a regression at 2am and need to remember which Redis command actually helps.

Install to Claude Code

npx -y skills add redis/agent-skills --skill redis-observability --agent claude-code