Date: 2022-09-23
Time: 09:30–10:20
Room: The Innovation Space
Level: Intermediate
Feedback: Leave feedback
Postgres provides a plethora of performance metrics useful for monitoring tools available through SQL. However a number of difficulties make deploying these tools a very manual process requiring a lot of custom tooling. Each site needs to arrange to deploy and configure agents and define the metrics, alerts and dashboards they need.
I have a new vision for the future of observability in Postgres where it integrates cleanly and automatically with widely used tools. Ultimately starting up a Postgres instance should trigger service discovery and metrics collection and enabling alerts and dashboards should be a simple matter of importing standard preconfigured packages into Prometheus, Grafana or other tools where they would "just work" for basic database monitoring with no customization needed.
This will require changes in core Postgres and contrib modules to adopt open standards for service discovery and metrics exporting.