Data Infrastructure

Runs GitLab’s datastores and data movement as internal services. Every system here has an owner who answers for its SLOs, capacity, cost, and upgrades.

Scope

We run GitLab’s datastores and data movement as internal services. Every system here has an owner who answers for its SLOs, capacity, cost, and upgrades. We don’t write product queries or business logic: the Event Platform writes into this layer, and Intelligence reads from it.

Teams


Data Storage
Runs ClickHouse and GitLab's analytics datastores as internal services.
Message Queue and Replication/CDC
Runs the pipes: NATS as the messaging backbone, and Siphon for replication and change data capture into the analytics stores.