What’s Breaking My Build? AIOps Tooling to Aid Your CI/CD Workflows
summary
An open-source AIOps toolkit called ai for ci collects TestGrid dashboards, Prow build logs, GitHub PR metadata and OpenShift telemetry, runs ~20 Jupyter notebooks to compute KPIs (flake severity, pass/fail rates), and stores metrics in Ceph S3 and Trino for Superset dashboards. Logs are cleaned and clustered with TF-IDF clustering to surface platform vs prep failures, a PR time-to-merge classifier bins outcomes into 10 ranges, pipelines run via Kubeflow pipelines and results are served with Seldon deployment (API returns 200).
AI-written summary · about
chapters (6)
- 0:00Start
- 4:11Our current initiatives
- 5:02Overview
- 6:22Open Data Sources
- 16:01Time to Merge of a PR
- 35:48Engage
related talks
26mAIOps for CI with Kubeflow Pipelines
31mTowards an open operations data community
34mOptimize Long Test Runtimes Using Open AIOps
29mFrom Flaky Builds to Self-Healing Deployments: An OTel-Driven CI/CD...
35mGaining Observability Into Your CI/CD Pipeline
35mBeyond the chatbot, beyond the IDE, where can AI help your CI/CD
25mMaximize CI/CD efficiency using AI and a vertically integrated stack
42mCI/CD - Let The Robots Do Their Job!