BI integration overview

Choose a guide below to connect a BI tool and check its results against a small, reproducible business dataset. A successful connection test alone does not establish that schema discovery, saved metrics, Decimal serialization and dashboards all work.

Prerequisites

Install MatrixOne using the standalone deployment guide and install the selected BI tool using its upstream instructions. Configure a database endpoint reachable from the BI service, normally the MatrixOne MySQL-protocol port 6001. In a container, localhost points to that container; use the MatrixOne service name or a reachable host address.

Prepare the shared dataset and KPI checks as an administrator. Use its read-only bi_reader account for BI connections. Replace MO_HOST, MO_USER and password placeholders with your deployment’s values; URI passwords must be percent-encoded. Choose the correct MatrixOne account-qualified username if using a non-default account.

Examples explicitly disabling TLS reproduce an isolated local test only. For remote connections, configure MatrixOne TLS and the tool’s CA/hostname verification; adjust the example driver options accordingly. TLS configuration was not covered by the compatibility run.

Versioned compatibility observations

The following observations were recorded on 2026-10-13, against MatrixOne commit 23d8f61252f3017dff218b530817a240206fd762, which reported 8.0.30-MatrixOne-v. The MySQL-compatible version string does not uniquely identify a MatrixOne release. These are historical observations of this particular binary and tool versions, not a certification of every MatrixOne release or every UI workflow. Retest your deployed combination before relying on a workaround or assuming an issue is resolved.

Tool

Tested version

Observed path

Metabase

v0.64.1.5

Saved SQL models, metrics and dashboard verified; sync limitations

Superset

6.1.0

SQL Lab / chart API blocked for some types; view reflection blocked

Grafana

13.2.3 (MySQL plugin 13.1.2)

Native datasource connection blocked in tested combination

Redash

26.9.0 / mysqlclient 2.1.1

Native runner verified; Decimal JSON precision limitation

DataEase

2.10.17 / Connector/J 8.2.0

Saved datasource and SQL preview API verified; preview row limit

Evidence

@evidence-dev/mysql 1.1.6

Native source connector blocked; diagnostic patch is not a setup option

Rill

0.90.5

Native ingestion and metrics verified; local engine / precision limitations

Apache Zeppelin

0.12.0 / Connector/J 8.4.0

Saved JDBC notebook execution verified; text formatting limitations

SQLPad

7.5.7

Native REST queries and saved SQL verified; legacy archived tool

Lightdash

2.573.1

No native MySQL warehouse adapter; no direct MatrixOne query

Follow FineBI, Yonghong BI or the existing Superset monitoring tutorial for their original workflows. The new Superset guide adds version-specific analytic checks without changing those historical instructions.

Validate an integration

  1. Run SELECT 1 and record both MatrixOne and driver versions.

  2. Run the shared total, category and daily queries; compare complete rows with the expected values, including nulls and Decimal scale.

  3. Save and reopen a query/model, filter category A, and verify net amount 130.000000.

  4. Inspect the SQL generated by the BI tool. Keep refunds at one row per order before joining; count distinct customers at the requested grouping level.

  5. Check preview/export limits and Decimal values after serialization. A chart that looks correct with small amounts can still lose precision for large amounts.

  6. Determine where aggregation runs: SQL pushdown to MatrixOne, or a local engine after extraction, as in Rill. Refresh extracted data before checking freshness.

Guides