Skip to content

Update ci runners service metrics definitions

Updates ci-runners SLI definitions:

  • adds new SLI - saas_runner_instant_capacity
  • replaces the previous saas_runner_queues SLI by a new one, using runner-sourced metric for calculating the apdex score. This will allow us to differentiate SLI of runner shards and even define different SLI for different times (like for example allow slightly longer queue durations during nights or weekends when we usually don't expect as big load as during middle of working day; in that cases we will require more time to scale up the ephemeral jobs and queue duration will by - by design - longer here).

Related to https://gitlab.com/gitlab-org/ci-cd/shared-runners/infrastructure/-/issues/194

Edited by Tomasz Maczukin

Merge request reports