Engineering Metrics
Some metrics to understand the engineering efficiency
Commit to Deploy Time
The time to take the code from commit to a deployment environment.
Measurement: Record the time elasted from code review to the code is merged.
Build Time
The time to build the changes (running unit test, integration test)
A good example: 10 mins build time will cost $9 per engineer who gets paid $54 per hour while waiting for build completion. Considering that two need to wait for the build five times a day, it will cost $450 per week and $21600 per year.
Measurement: Time to set up, tear down the testing environment, and run the test.
Queue Time
It depends on the team size and the throughput of the CI/CD pipeline. The bigger the team size is, the more build requests send to pipeline, and it leads more queue time for the engineers to wait for the build to complete, and test their changes.
Measurement: Track the length of waiting time before the next job is kicked off.
Frequency of a broken production branch
It is crucial to ensure the pipeline and deployment environment is in a deployable state so that the engineers can continue shipping software to production.
Measurement: Total length of broken time / Total working time
Engineering Overhead
It includes headcount, expense on the license, managed services, maintenance on tools.
Measurement: Amount of time spend on tooling or process, instead of building features.