Narrative

Having metrics for a Keycloak installation is essential to identify bottlenecks ahead of time, and do trouble shooting during incidents. Metrics are the foundation for service level indicators ("measure how well my system is running") / service level objectives ("this is what I promised to my consumers"). 

Providing dashboards is a follow-up task which will build on top of these metrics.

Value proposition

• Help admins to run a stable Keycloak environment for trouble shooting, capacity planning and negotiating service levels

Goal

Provide documentation and examples for:

• Metrics that an admin can use as service level indicators (SLI) and a base for a Service Level Objectiveas those metrics capture the observed behavior of the system by a user. Those SLIs are usually negotiated with other stakeholders of the system.
• Metrics that an admin of Keycloak can use for trouble shooting, finding root causes of errors, slowness of the system, and looming problems due to resource shortages for example in connection pools or memory usage.
• Metrics that an admin can use for capacity planning of the system, especially those we use in the sizing guide (number of password based logins, number token refreshes, number of client credential grants). A growth in those metrics would indicate they would need to increase the resources of Keycloak in-line with actual and anticipated growth.  

Non-Goal

• (future) Dashboards
• (future) Error Budgets

Implementation Note

• Documentation published as guides in the Keycloak docs
• Add Keycloak metrics as necessary for the above goals.