Map Every Apache Druid Datasource on One Canvas

Druid answers OLAP queries in under a second from time-partitioned segments. DbSchema connects to the Broker over the Avatica JDBC protocol and diagrams the datasources behind those queries — columns, types, and structure on one canvas.

Every datasource with its columns and types, drawn as a diagram
Check what rollup produced right after an ingestion run
Compose TIME_FLOOR and approximate aggregations with completion
Sketch new datasource schemas before touching an ingestion spec

available for

ISO-27001 certified

Free Community Edition No signup required 15-day Architect trial included

Free Download

DbSchema mapping Apache Druid datasources as schema diagrams

Trusted by teams at

A Dozen Ingestion Specs Define Your Model
And No Two People Read It the Same

In Druid, each datasource's shape is buried in JSON — a dimensionsSpec here, a metricsSpec there, rollup changing what actually lands. Analysts see only the query side and guess at the rest.

DbSchema introspects the Broker's catalog and draws the result: real datasources, real columns, real types — the model as it exists, not as the spec intended.

Map, Query & Explain
Your Druid Datasources

The analytics model — drawn, queried, and documented from one desktop app.

Group and annotate Apache Druid datasources in DbSchema layouts

Group Datasources by Subject Area

Arrange datasources into annotated layouts per pipeline or domain, so the analytics estate reads as a map instead of a flat list of names.

Write Druid SQL with TIME_FLOOR in the DbSchema editor

Draft Druid SQL With Completion

Compose TIME_FLOOR() buckets and APPROX_COUNT_DISTINCT() aggregations with datasource and column completion, then hand the verified query to your BI tool.

HTML documentation of Apache Druid datasources from DbSchema

Explain the Pipeline to Analysts

Publish documentation that captures column names, types, and the relationships modeled in your ingestion specs — readable without ever opening a spec file.

Re-Ingesting Costs Hours
Redrawing a Diagram Costs Minutes

The design model is a local file, so a datasource can be drafted, discussed, and corrected on the canvas before the first segment is ever built.

Draft Apache Druid datasource schemas offline

Draft on the Canvas

Lay out columns and types for a new datasource before writing its ingestion spec.

Catch Druid schema mistakes before rollup

Fail on Paper, Not in Rollup

A wrong column caught in review costs a redraw — not another ingestion cycle.

Version the Druid schema model with ingestion specs

Version It With the Specs

Commit the model next to your ingestion JSON and diff schema changes over time.

Validate Druid ingestion results in DbSchema

Verify After Ingestion

Sample fresh rows in the data explorer and compare against what the spec promised.

Connect Through the Broker via Avatica

Druid speaks JDBC through Apache Calcite Avatica: load the Avatica standalone JAR — see the Druid JDBC driver page — into DbSchema's driver manager and use the remote driver class org.apache.calcite.avatica.remote.Driver. The URL points at the Broker node: jdbc:avatica:remote:url=http://broker:8082/druid/v2/sql/avatica/, with 8082 as the default Broker port. On secured clusters, switch the scheme to https and enter credentials in the connection dialog.