Preventing catastrophes using OpenShift data
summary
Red Hat's Connected Customer Experience team harvests telemetry since OpenShift 4.2: prometheus metrics and alerts every five minutes via Prometheus metrics, and structured JSON archives every two hours via the Insights operator. They detect concrete regressions — e.g. a spike in kube api errors tied to 4.5.16 that later disappears in 4.6/4.5.9 — and run automated reports with Argo Workflows and dashboards in Superset to prioritize fixes and confirm remediation.
AI-written summary · about
chapters (9)
- 0:00Start
- 7:04Health Data Collection
- 8:11Insights Operator
- 9:45Examples of Error Messages from Various Openshift Components
- 13:28Openshift Users
- 15:33Jupiter Notebooks
- 17:33Argo Workflows
- 19:44Report Examples
- 25:05Filtering the Data
related talks
24mUncovering Community and Project Insights Through Data Driven Methods
32mScalable Monitoring Using Prometheus with Apache Spark
33mAccidental DataOps: How we matured our organization with data, and how you can too!
31mTowards an open operations data community
20mData Incidents - When things go wrong
21mChallenges, Best Practices, and Solutions for Monitoring and Alerting...
35mBuilding a Real Time Metrics Database
19mConfident Data Migration: automatic regression testing