Observability built for the ClickHouse® database

CHOps gives you deep visibility into ClickHouse® clusters without a separate metrics pipeline. It reads directly from system tables, so you get real-time dashboards for CPU, memory, disk, and network with almost no extra storage.

For slow query analysis, the query profiler shows a flame graph of where a query spent its time, the processors profile shows it as a pipeline, and query metrics break down CPU, memory, and IO second by second. When something breaks, DVR-style playback lets you rewind to the exact moment instead of reconstructing it from logs.

Merges, mutations, parts, queues, and replica status are all in one place, so you catch problems like a rising part count or a stuck mutation before they turn into incidents. Open source, no agents on your database nodes.