We wrote a guide + demo showing a code‑first path from Postgres to ClickHouse using Debezium and Redpanda, with MooseStack as the dx/glue layer.
This post targets teams who want code ownership: typed streams, stateless streaming transforms, and OLAP tables defined in code, with a local dev server to spin up ClickHouse + Redpanda + Debezium + Postgres in one command.
Blog is linked in the repo.
Right now the demo solely focuses on the local dev experience, looking for community input regarding best practices for running Debezium in production.
If your new APIs use the same endpoint names and the same request/response schema (same field names, types, shape), then swapping the base URL in your fetch calls could get you most of the way there without touching your visualization components.
In practice I get that there’s usually some drift, like maybe a field is renamed or you have different error handling. Depending on how strictly your front-end components assume the old schema, you’ll need to update those areas.
This post targets teams who want code ownership: typed streams, stateless streaming transforms, and OLAP tables defined in code, with a local dev server to spin up ClickHouse + Redpanda + Debezium + Postgres in one command.
Blog is linked in the repo.
Right now the demo solely focuses on the local dev experience, looking for community input regarding best practices for running Debezium in production.