Stop delivering crap to your customer without knowing!
“Huh, when did we introduce these code violations?” or “When did we increase our test coverage?”
Are these questions often asked by members of your team? Chances are that you are in an area where you are using CI/CD solutions but insight into the data produced can be upgraded.
Companies often already made a start with CI/CD: running the tools, doing the checks, and getting metrics. In a pipeline, this is a snapshot in time. A team can use quality gates on these metrics to define when a job is failed or is successful. How great would it be if these metrics and quality gates could be visualized on a dashboard?
An API for quality gates
That’s exactly what we did. We’ve created an API where we can post our standardized quality gates. Next to the metric and the threshold, we store in which pipeline it occurred and on what commit and branch. The upload to the API is added to our CI/CD tool suite and deployed in all our pipelines.
We created multiple views of the dashboard in Grafana. We can show the trend graphs of Quality gates over time next to their thresholds. Or we show the current amounts next to a traffic light visualizing a threshold breach.
As the metrics are already standardized, we could also standardize the dashboards. Due to the standardization, we can deploy this for every project. So, every internal project within Capgemini Engineering Netherlands can use this API and the predefined dashboards to their benefit, without configuring, deploying, changing, or doing anything. If they use the pipeline examples with our tooling provided by us.
In one of my previous projects, we displayed the dashboards on our wall. It started a game between some developers to decrease the number of warnings. But also started discussions between team members, colleagues, and even visitors on why certain metrics changed or what they meant.
Want to know more about quality gates and how to get metrics of your CI/CD tool suite? Wait for my next blog or if you can’t wait, contact us. But in the meantime, start gathering your metrics; with them, your improvements become empirical.
