Engineering for Reliability
Migrating to the managed service for Prometheus
Twenty-three substantive episodes on SLOs, telemetry, alerting, investigation, and operations
Twenty-three 2021–2022 episodes connect user-facing reliability targets to metrics, logs, traces, and alerts. Examples show instrumentation, Prometheus collection, quota monitoring, GKE investigation, and observability cost controls.
Google Cloud Tech 21 of 23
In this collection Browse 23 summaries 21 of 23
Prometheus collects measurements by reading, or scraping, an application's metrics endpoint. Yuri, an SRE at Google Cloud, demonstrates moving one application's collection configuration to Google's managed service so the team no longer has to operate its own Prometheus infrastructure 00:00–02:11(opens in a new tab).
Key Points Covered
- Establish the working configuration first: Prometheus Operator manages the existing Prometheus deployment. A
ServiceMonitorresource tells it to discover and scrape the application's/metricsendpoint; the presenter confirms that the app is a target and that its measurements are arriving 00:00–02:11(opens in a new tab). - Translate the collection configuration: After enabling managed collection, the walkthrough replaces the
ServiceMonitorwith aPodMonitoringresource. The configuration is similar, but the presenter explicitly warns that some fields may need to change 02:11–03:03(opens in a new tab). - Check data, not just successful configuration: Queries in PromQL, Prometheus's query language, and Metrics Explorer confirm that the application's measurements arrive after migration. This validates the example's collection path, not complete compatibility for every Prometheus feature or configuration 03:03–03:55(opens in a new tab).
Full video: https://www.youtube.com/watch?v=YjEh_V7MRc4(opens in a new tab)