// the find
ongres/stackgres
StackGres Operator, Full Stack PostgreSQL on Kubernetes // !! Mirror repository of https://gitlab.com/ongresinc/stackgres, only accept Merge Requests there.
StackGres is a Kubernetes operator that packages PostgreSQL with the whole operational stack around it — Patroni for failover, connection pooling, backups, Prometheus/Grafana monitoring, and a web console — as one deployable unit. It's for teams running Postgres on Kubernetes who want more than a bare operator and don't want to hand-wire Patroni, pgBouncer, and monitoring themselves.
The operator is built on Quarkus and compiled to a native GraalVM binary, so you get genuinely fast pod startup and low memory overhead compared to typical JVM operators. HA failover rides on Patroni without needing an external DCS like etcd or Consul — it talks to the Kubernetes API directly, which removes a whole class of infrastructure to babysit. The feature list isn't just marketing: distributed logging, Envoy-based query observability, and server-side connection pooling are real, non-trivial components to get working together, and they ship integrated rather than as 'bring your own.'
It's AGPLv3, which is a real adoption blocker for any company that might want to modify it internally without open-sourcing changes — they sell a commercial license specifically to work around this, which tells you how often it comes up. Development actually happens on GitLab; this GitHub repo is a read-only mirror, so GitHub issues and PRs go nowhere and you have to context-switch platforms to contribute or even file a bug. Layering a Java/Quarkus operator on top of Patroni adds a second control plane to reason about when something goes wrong with failover — more moving parts than operators that let Patroni run closer to bare. Adoption is modest for the surface area covered (1.4k stars, 76 forks for an 'enterprise Postgres distribution'), so you're betting on a comparatively small community versus something like CloudNativePG or Zalando's operator.