Having accurate build metrics allows us greater insight into how changes we make to infrastructure, configuration, and tooling impact both the stability and performance of our CI infrastructure. Metrics should be regularly collected from Jenkins and aggregated for consumption by some existing analyzer/visualizer like Grafana.
Metrics should be useful enough to help us assess:
- How changes to job configurations impact periodic build duration averages
- How changes to CI tooling impact periodic build duration averages
- How changes to infrastructure and underlying resources impact periodic build durations averages on each node
- Capacity planning (executor availability and usage), both current utilization and trends
- Jenkins master resource utilization trends