Clickhouse Managed Postgres Rca

When your ClickHouse-managed Postgres instance starts crawling, this skill runs a six-step diagnostic workflow that hits two Beta APIs: Prometheus for system gauges like cache hit ratio and active connections, and the Slow Query Patterns endpoint for per-digest latency and IO stats. It triangulates evidence using three built-in heuristics (full scan, N+1 loop, write congestion), then writes up a root cause report with a recommended fix but won't touch your database. The workflow is strict about not hardcoding field names since both APIs are still Beta, so it fetches the OpenAPI spec fresh every session. Nice safety rails for an observability layer that's still moving.

Install to Claude Code

npx -y skills add clickhouse/agent-skills --skill clickhouse-managed-postgres-rca --agent claude-code