Devconf.cz 2023

In this era of automation, continuous integration and delivery(CI/CD) with code hosting platforms has played a vital role in software development workflows. A huge number of contributions are made to Open Source projects and are subjected to automated tests/builds before being merged. However, some tests run for longer durations than expected and are often painful as they block the CI/CD process.
In this session, we will introduce an open AIOps tool called AI4CI that consists of a set of Jupyter notebooks automated using Kubeflow pipelines and collects data from K8s CI/CD platforms, analyze and visualize the KPI metrics, develop and deploy ML models using Openshift.
By analyzing the test failure times, we aim to predict an Optimal Stopping Point for a test after which the test is most likely going to fail and better support in managing the development resources efficiently. The attendees will learn about how the Optimal Stopping Point model can resolve the bottlenecks in CI/CD systems.