See where your disk actually goes
Tables ranked by size, compression ratios per column, parts and partitions — so you know what to drop, TTL or re-encode before the disk fills.
The tables eating your cluster, ranked
Every table with its size on disk, row count and compression ratio, sorted so the expensive ones are at the top. The answer to "why is the disk full" is the first row, not a query you have to write.
- Compressed and uncompressed size side by side
- Compression ratio per table and per column
- Rows, parts and partitions per table
Watch headroom, not just usage
Per-disk usage with what is left, next to the merge and parts pressure that decides how fast it moves. Catch the trend while there is still room to act on it.
- Free space per disk and per volume
- Merge and mutation pressure alongside usage
- Health checks that alert before a disk fills
More of Storage in action
Everything in the box
Table sizes
Compressed, uncompressed and row counts.
Compression
Ratio per table and per column.
Parts
Parts and partitions broken out per table.
Disks
Usage and free space per disk and volume.
Alerting
Health checks fire before a disk fills.
AI advice
Ask the agent what to TTL, drop or re-encode.
Common questions
Where do the sizes come from?
system.parts and system.columns — the same numbers ClickHouse reports, aggregated per table, column and partition.
Can it tell me what to delete?
The AI agent reads the same storage tools and will recommend TTLs, codecs or drops — recommend-only, nothing is applied without you.
Does it work on a cluster?
Yes. Sizes are reported per host, and the host switcher moves between clusters.
More chmonitor features
Start monitoring — advisor and alerts built in
Open the hosted dashboard and connect a cluster, or self-host with Docker, Kubernetes, or from source. Self-hosting is always free.