You need to sign in or sign up before continuing.
Remove metrics that take long time from Service Ping generation for GitLab.com
Problem
Service Ping generation takes more than 40 hours in the last 2-3 weeks.
There is a risk to not be able to generate it successfully
Proposal
Remove/Optimize metrics that take long time from Service Ping generation for GitLab.com
Proposed metrics to be removed
See monitoring issue https://gitlab.com/gitlab-org/gitlab/-/issues/362739#note_975137024
-
counts.ci_internal_pipelines
12.5
hours - counts.issues_created_gitlab_alerts 3.47796 hours
- counts.issues_created_manually_from_alerts 3.47796
Next steps for this
-
Remove the metrics for GitLab.com -
Regenerate Service Ping when the changes are in production. This will require access to the production rails console. Follow https://gitlab.com/gitlab-org/gitlab/-/issues/362739#how-to-generate-usage-ping-for-gtilabcom -
Create individual issues for the broken metrics to fix them
Edited by Alina Mihaila