Understand how EDB packages and distributes ClickHouse, what's covered in this documentation, and what falls outside its scope.
How EDB provides ClickHouse
EDB Postgres AI for ClickHouse isn't a fork. EDB builds and distributes the open-source ClickHouse binary, identical to the upstream release, repackaged as RPMs, container images, and a Kubernetes operator for enterprise Linux distributions, and supported by EDB.
EDB's contribution is:
- Building the official open-source ClickHouse source at a pinned upstream release
- Packaging the binary as RPMs, a container image, and a Kubernetes operator, distributed through the EDB package repository
- Integrating ClickHouse into the EDB Postgres AI Analytics pillar alongside WarehousePG and the EDB Postgres Lakehouse
- Providing EDB support
The ClickHouse query engine, SQL dialect, table engines, and performance characteristics are identical to the upstream open-source release.
Verify you're running an EDB-packaged build:
SELECT value FROM system.build_options WHERE name = 'VERSION_OFFICIAL';
The output includes (EDB Build) for EDB-packaged releases.
What's included
EDB packages and supports the open-source ClickHouse binary. The following capabilities are available in the EDB distribution.
| Area | What's covered |
|---|---|
| Table engines | MergeTree family, integration engines (PostgreSQL, S3, Kafka, IcebergS3, DeltaLake, MongoDB, Redis, and more), and special engines (Distributed, MaterializedView, Dictionary, and more) |
| Database engines | Atomic, Replicated, PostgreSQL, MySQL, SQLite, DataLakeCatalog |
| SQL and query capabilities | Aggregations, window functions, CTEs, subqueries, materialized views, all major data formats (Parquet, JSON, CSV, Avro, ORC, Arrow), dictionaries, projections, and RBAC |
| Data lake integration | Read access to Apache Iceberg® tables via DataLakeCatalog (REST, AWS Glue, Databricks Unity Catalog, Hive, OneLake) and directly via IcebergS3, IcebergAzureBlobStorage, and related engines; Delta Lake and Apache Hudi via dedicated table engines |
| Operations | ClickHouse Keeper, clickhouse-client, clickhouse-local, and Named Collections |
What's not covered
The following features are exclusive to ClickHouse® Cloud, a separate commercial service from the upstream ClickHouse project. They're not part of the open-source binary and aren't covered in this documentation.
| Feature | Description |
|---|---|
| ClickPipes | Managed ingestion pipelines for Kafka, S3, Postgres CDC, MySQL, MongoDB, and others |
| SharedMergeTree | Proprietary table engine for ClickHouse® Cloud's shared-storage architecture |
| SQL Console | Web-based query editor with dashboards, visualizations, and GenAI SQL assistance |
| Query Insights | Managed query log analytics built into the Cloud console |
| Managed Postgres | Fully managed Postgres service collocated with ClickHouse |
| Compute-compute separation | Architecture for scaling compute independently of storage |
| Auto-scaling | Automatic vertical and horizontal scaling |
| Managed backups | Automated, configurable backups with external export |
| BYOC | Bring Your Own Cloud, a managed service deployed in your AWS account |
| Compliance certifications | HIPAA, SOC2, PCI DSS, ISO 27001 (managed by the Cloud service) |
| SSO / CMEK | Single sign-on and customer-managed encryption keys via Cloud console |