FY21-Q1 Monitor Engineering Manager OKR: Dogfood Monitor Metrics Dashboards

Key Results

• Replace SLA Dashboard from Grafana to use GitLab's metrics dashboard => 80%
  • 80% because we cloned it into https://gitlab.com/gitlab-com/dashboards-gitlab-com/-/environments/1790496/metrics?dashboard=.gitlab%2Fdashboards%2Fsla-dashboard.yml but have not replaced the Grafana dashboard yet

Notes

This relates to the Ops OKR: #6207 (closed)

---

Retrospection

Good

• Made a lot of progress in a quarter compared to the previous quarter
• Setup a recurring sync meeting with several stakeholders of the Infrastructure team
• More rapid feedback loop after getting Thanos to work with gitlab.com (since gitlab.com is deployed continuously and ops.gitlab.net is deployed monthly)
• Jsonnet made it easier to migrate over to our YML definitions for dashboards

Improve

• Handful of regressions that took place that made it difficult to make progress (gitlab-org/gitlab#217173 (closed), gitlab-org/gitlab#217618 (closed)). We need to improve our test coverage
• Mixed opinions from Infrastructure team in regards to which dashboards we could clone and/or replace. Mixed understanding among different GitLab team members about when we can start dogfooding certain dashboards. Sid helping clarify this in the last Infrastructure GC should help mitigate this and help with momentum.
• Andrew was the primary maintainer for runbooks MR. It is nice to have him involved but it has been a slight bottleneck at getting faster iteration and revisions for the metrics dashboard on the runbooks project