// the find
drasi-project/drasi-platform
The Data Change Processing platform
Drasi is a CDC-style event processing platform that lets you write Cypher queries against live data sources and fire reactions when query result sets change. It sits between your data stores and your automation layer — think triggers that work across systems, not just within one database. CNCF Sandbox project, primarily Kubernetes-targeted, with sources for Postgres, Cosmos DB, and a few others.
Using Cypher for continuous queries is a genuinely good idea — graph query semantics map well to the 'join events across sources' problem that's painful to express in SQL or stream processing DSLs. The architecture cleanly separates Sources, Queries, and Reactions into independent deployment units, so you can scale or replace each without touching the others. The tooling story is solid for a sandbox project: a CLI, a VS Code extension with schema validation, and a debug reaction for live query inspection. Control plane is Rust, which is the right call for something managing Kubernetes CRDs and Redis streams.
Kubernetes-only right now — there's no local-first or Docker Compose path that's officially supported, which makes the getting-started experience painful if you just want to evaluate it. The README says 'early release' and the source connector catalog is thin; if your data isn't in Postgres, Cosmos DB, or one of a handful of others, you're writing your own source provider before you can do anything useful. Cypher is expressive but it's also a query language most teams don't already know, so the learning curve compounds with the operational complexity of running a new platform. No mention of backpressure or at-least-once vs exactly-once delivery guarantees in the surface-level docs, which matters a lot before you trust this with production side effects.