Cloud-Native Postgres Observability: From Client Apps to Underlying Cloud Resources

What does it take to maintain observability over PostgreSQL databases in cloud environments where systems are constantly shifting? This session explores how to keep visibility intact across complex infrastructure where databases migrate between nodes and storage performance changes on the fly.

We walk through how eBPF enables detailed service mapping and instruments Postgres calls from any client application to a specific database, giving teams a clear, real-time picture of their data layer across a distributed system. The session also covers automatic topology discovery within High Availability Postgres clusters managed by various Kubernetes Operators, and how to monitor the underlying cloud resources, including virtual machines, storage volumes, and network components, that Postgres performance depends on.

If you're responsible for PostgreSQL observability in a Kubernetes or cloud-native environment, this session offers a concrete look at how to maintain visibility without losing your mind to complexity.