This walks you through setting up TimescaleDB hypertables with all the options that actually matter: partition columns, segment_by for grouping hot query paths, order_by for compression, and sparse indexes to skip decompression. It's built for insert-heavy patterns like metrics, IoT sensors, event logs, or transaction records. The segment_by guidance is especially useful since most people either skip it or add too many columns. One thing to know: it assumes you want columnstore compression by default, which you probably do unless you're using pgvector. The examples are concrete enough that you can adapt them without reading TimescaleDB docs for the third time this month.
npx -y skills add timescale/pg-aiguide --skill setup-timescaledb-hypertables --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
juliusbrussee/caveman
mattpocock/skills
shadcn/improve
obra/superpowers
forrestchang/andrej-karpathy-skills
vercel-labs/skills