GitLab

Simplify service aggregated apdex calculation

See merge request gitlab-com/runbooks!3111

Remove handler label from registry dashboards

See merge request gitlab-com/runbooks!3110

Add kube resources to monitoring service

See merge request gitlab-com/runbooks!3096

refactor(metrics): preparation for aggregation sets

See merge request gitlab-com/runbooks!3106

Rearranges recording rules into aggregation sets to reduce churn in upcoming MR

Add a filters into stage group dashboards

Closes gitlab-com/gl-infra/scalability#740

See merge request gitlab-com/runbooks!3100

Add diagnostic links for Cloudflare

See merge request gitlab-com/runbooks!3108

Corrective Action: gitlab-com/gl-infra/production#3317 - find project with pages domain

See merge request gitlab-com/runbooks!3104

Document gitlab-org/grafana-dashboards as archive for deleted dashboards

See merge request gitlab-com/runbooks!3107

SLIs for thanos memcached

See merge request gitlab-com/runbooks!3097

Fix registry logging links

See merge request gitlab-com/runbooks!3102

Rethink chef-client staleness alerts

See merge request gitlab-com/runbooks!3101

Only alert on stale chef client runs on instances chef-client is enabled
on.

Add an alert for when chef client has been disabled for over 24h, much
longer than the 5h staleness alert.

This allows us to distinguish between unexpected stale runs and
intentional pauses, while alerting when a pause might have been
accidentally left in place.

For example, this scenario arises every time the CI runner manager fleet
is updated:
https://gitlab.com/gitlab-com/runbooks/-/blob/master/docs/runner/update-gitlab-runner-on-managers.md#procedure-description.

Remove the duplicate/alternative ChefClientStale alert for simplicity.
That alert had an incident_project configured: move it to the now-only
ChefClientStale alert, and change the incident_project to production
from infrastructure. Infrastructure issue alerts are harder to discover
than production issue alerts, because Woodhouse posts to
`#incident-management` with news of new production issue alerts but not
Infrastructure ones. This change effectively makes the alert slightly
noisier: on-calls often pay closer attention to incident-management than
the non-paging alerts channel.

By error, then by staleness

Update container registry HTTP metrics labels

See merge request gitlab-com/runbooks!3094

Corrects links to moved documentation for Redis

Closes #57

See merge request gitlab-com/runbooks!3098

Add schema for rails.api logs

See merge request gitlab-com/runbooks!2886

Add feature_flag.json

Closes gitlab-com/gl-infra/production#3130

See merge request gitlab-com/runbooks!3012

Sync grafana and kibana time ranges

Closes gitlab-com/gl-infra/scalability#769

See merge request gitlab-com/runbooks!3095

add script for finding duplicate repos

See merge request gitlab-com/runbooks!2784