Show violation counts per Rails endpoint in a table
In a discussion on Slack with @kniechajewicz it was brought up that it is currently not possible to see the worst behaving Rails endpoints for the full 28 day period we use for error budgets for stage groups. We do have it in the logs for the past 7 days in a slow loading dashboard. Example for ~"group::acquisition".
When we've included the error information in GitLab-SLI metrics and aggregated them globally, we could build the same thing in a Grafana table. Showing the overall request count, apdex violation count and error count in a table.
We could add this in the SLI details:
collapsed row and use each significant_label
of an SLI as an aggregation label in the queries: this means we'll have a column per significant label.
We can start with this on the existing dashboard, but perhaps it would be interesting to build an SLI exploration dashboard for all SLIs that allows more filtering. This could then replace the Controller dashboard we currently have per service.