Skip to main content

Database Explorer

Browse your ClickHouse databases, tables, parts, and columns with detailed metadata and visualizations.

Database Tree​

The left panel shows a hierarchical tree of all databases and tables. Click any table to see its detail view.

Table Detail​

For each table, you can inspect:

  • Schema - Column names, types, compression codecs, and sizes
  • Parts - Active data parts with size, rows, level, and partition info
  • Part Inspector - Inspect individual parts with column-level statistics
  • Merge Timeline - Historical merge activity for the table
  • Lineage - Visual merge lineage tree showing how parts were created and merged

Parts Visualization​

The 3D parts view is a learning tool for understanding how ClickHouse's MergeTree engine works. Each data part is rendered as a block, sized proportionally to its data volume. Colors indicate:

  • Part level - how many merges a part has been through, from freshly inserted (level 0) up through successive merge rounds
  • Partition membership - which partition each part belongs to
  • Active vs. inactive state - whether a part is still live or has been replaced by a merge result

By watching parts appear from inserts and gradually consolidate through merges, you can build an intuition for how the merge algorithm selects parts to combine - typically picking groups of similarly-sized parts within the same partition to keep the tree balanced and write amplification low.

Merge Lineage​

The lineage visualization traces the full lifecycle of parts over time:

  1. New parts created by INSERTs arrive at level 0
  2. The merge algorithm selects groups of parts and combines them into larger, higher-level parts
  3. Mutations rewrite parts in place, creating new versions
  4. The resulting hierarchy shows the progression from many small parts to fewer large ones

This makes it easy to see how ClickHouse decides what to merge and when, and to spot problems like:

  • Parts not merging (too many small parts accumulating)
  • Unbalanced merge trees (some partitions merging faster than others)
  • Stuck mutations blocking normal merge progress