How engineering leaders can have more visibility on their GitHub Actions

Engineering leaders are responsible for development processes and their teams experience. Todays leaders in engineering must take decisions based on data. What measurements ought to be monitored as the key performance indicators for the team on their CI/CD process? Which specific facts should technical leaders focus on?

Serkan Ozal, CTO, and Co-Founder at Thundra says

Without a proper CI pipeline, it is impossible to send your code bug-free to production. If the failed tests running on your CI are not investigated properly, it is expected that the code will break something in production.

Common things when developers see a failed test or workflow are:

• Re-running the whole workflow from the start
• Try to reproduce the bug on local
• Dig into the log piles

Although the traditional ways are still helping the developers, it loses so much time than expected.

Lemi Orhan, CTO, and Co-Founder at Craftgate mention how crucial is to be aware of the CI performance at a high development pace.

We have more than 160 workflow runs each day. It is very hard to identify when and why our workflows have failed. It is likely that one of our workflows start to run 5 times slower than usual. We needed to be alerted in some way.

Metrics about your CI and test performance attempt to guide you to a defect-free and quick software delivery when you take a broad view of the process of shipping software from development to production.

Stian Thorgensen, Keycloak Project Lead shares some insight about why they chose GitHub Actions over many alternatives

As an open-source project, we were using Travis CI, but we switched to GitHub Actions very quickly. It allows you to have twice as many parallel runs, for free. Thats why we have switched to it.

However, GitHub Actions fall short of answering crucial questions like;

Which tests are blocking the PR?

Which changed lines havent been tested?

What is the most costly workflow?

Is my workflow success rate decreasing?

Solution: Foresight

Foresight brings visibility into CI workflows and test runs, enabling developers to debug CI, build, and test failures and latencies quickly. Furthermore, Foresights change impact analysis helps you analyze the impact of the code changes to detect defects early in the release cycle.

With the help of the Highlights dashboard, engineering managers, and CTOs are eligible to answer the below questions;

• Which workflows fail frequently and need to be optimized?
• What is the duration/success/cost trend in my workflows?
• Where do we need to start taking action about health and optimization?

Stian tells his experience about Foresight until now:

Weve been avoiding non-releasable junks due to the broken, unstable, and flaky tests. Foresight gives you the ability to look at more than one workflow at a time. You can easily see the tests and the results without losing your way in the log piles.

You may check the new codes level of testing and how it affects your test results using Foresights PR view. The changes that prolong the merging process are easy to spot using Foresight. As a result, Foresight assists you in maintaining your contributions to your open-source project in a secure and effective way.

Lemi adds to his initial experiences:

We have many repositories including several mono repositories. Foresight helped us to monitor all the workflows, runs, and test results from one single place, to improve the pipeline by the alerts, and to save a lot of time. I highly recommend it to the community.

If you would like to watch the Foresight demo and hear more from Serkan, Lemi and Stian, dont forget to check out the webinar recording

Create your free Foresight account and start monitoring your GitHub CI pipelines and tests. You can reach out to the Foresight team via its Discord community.