chmonitor's product design system + UX conventions, so every new feature looks and behaves consistent with the rest of the dashboard. Use BEFORE building or reviewing any UI: new pages, charts, cards, dialogs, empty/error/loading states, badges, onboarding,…
Work on chmonitor's Cloud (SaaS) vs self-hosted (OSS) behaviour from ONE codebase. Use when changing how dash.chmonitor.dev differs from Docker/K8s/OSS builds: the cloud-mode flag, public read-only demo hosts, the welcome/setup onboarding page, per-user (D1)…
Analyze cluster data and metrics using raw SQL recipes against system.query_log and system.parts when no dedicated tool exists.
Structured triage recipes for common ClickHouse incidents: disk, errors, replication, mutations, cluster health, and slow queries.
Recommend table ORDER BY keys, partition strategies, column data-type right-sizing, codecs, skip indexes, and projections for ClickHouse tables.
Exact column names for the system tables the agent queries most (processes, query_log, parts, merges, mutations, replicas, replication_queue, disks, settings, zookeeper, users/grants, metrics) plus rules for choosing dedicated tools over raw SQL. Load before…
Schema migrations: ALTER patterns, engine changes, zero-downtime swaps, clickhouse-local offline migrations, lightweight UPDATE/DELETE strategies, and Postgres→ClickHouse migration planning (type mapping, schema pitfalls, PeerDB CDC, validation, schema…
Advanced query tuning: join algorithms, skip index selection, EXPLAIN interpretation, ProfileEvents profiling, and optimizer settings. Also the autonomous diagnose-loop for open-ended questions like why is my database slow?