If you need to run PostgreSQL within Kubernetes environments, there are too many options. This session aims to assist in selecting the right operator by sharing the author's findings and production experiences. Key considerations include operator capabilities, security practices, extension support, connection pooling, and observability. It highlights the importance of self-service for developers, production readiness, performance predictability, operational maturity, and autoscaling. The most popular projects are evaluated:
DIY solutions as Baseline
Bitnami Helm Chart
Zalando Operator, Cybertec PG-operator
Postgres Operator from CrunchyData, Percona Operator
CloudNative PG
StackGres
The session will answer a lot - which operator can corrupt or lose my data? Which operator can really do seamless failover in a fully automated fashion? How easily can the developers provision their snapshot of the production backup with an operator? Will there be a clear best PG operator winner in 2024?