Debezium Platform Lab¶
Build a data pipeline that carries every change in your database — every insert, update and delete — to wherever you need it, configured in a web UI instead of hand-written connector plumbing.
This lab stands up the Debezium Platform, either empty, ready for you to explore and configure, or with a seeded demo database and a working pipeline, all with a couple simple commands.
Community project
Not affiliated with, or endorsed by, the Debezium project or Red Hat.
What you actually get¶
A running platform where you define pipelines by filling in forms:
- Sources — a database you want to capture from. PostgreSQL, MongoDB, Oracle, SQL Server, and more; the platform's catalogue keeps growing.
- Destinations — where change events are delivered. An HTTP endpoint, Kafka, and other sinks the platform supports.
- Transforms — optional reshaping between the two.
- Pipelines — a source bound to a destination, running and observable.
Everything the lab creates through the UI is also expressible as JSON, so once a pipeline works, you can commit it and recreate it anywhere. Nothing you build here is trapped in a click-path.
The default is the platform plus monitoring — Prometheus, Grafana, the otel operator, ingress-nginx, ready for you to explore and configure. You can also start with a seeded demo database and a working pipeline delivering change events to a plain HTTP endpoint, because that makes them immediately visible: change a row, watch the JSON arrive. Kafka is one configuration change away when you want it — swapping the destination is the point of the platform, not a migration.
Do I need Kubernetes?¶
For now, yes — the lab deploys onto Kubernetes, locally with Kind or onto a cluster you already run. just recipes controlled by environment variables abstract you from raw helm/kubernetes commands:
and your platform is up-and-running with monitoring and observability built in with sensible defaults.
Planned: a host-based path — the same pipelines on a machine with a container runtime, no Kubernetes involved.
The work that makes this possible is the Debezium team's: Debezium 3.7.0.Alpha1 added a native distribution module for Debezium Server: Debezium Server can now be built as a native executable, with a faster startup and a smaller memory footprint than the JVM distribution.
To stay informed, subscribe to the Debezium blog.
Where to go¶
| If you want to | Go to |
|---|---|
| Get it running now | Get running |
| Fix something that failed | Troubleshooting |
| Deploy a more complex flow | Guides |
Before you start¶
How much resources you need depends on where the platform runs.
| Deployment | Runs on | Needs |
|---|---|---|
| Kind | your machine | roughly 2 CPUs, 4 GB RAM, 10 GB free disk |
| k3s | a server you connect to | the same, on that server — your own machine only needs the tools |
| Existing cluster | wherever it already runs | capacity on the cluster; nothing in particular here |