How much information is in an empty list?
summary
Red Hat builds proactive recommendations for OpenShift by shipping an insights operator that collects cluster snapshots every two hours, zips and sanitizes them, and sends them to a cloud recommendation service that flags fixes and prevents roughly hundreds of support cases a month. Empty or missing items in the archive can mean a true absent config map, a version of the operator that didn’t collect that artifact, archive size limits, or collection errors so rules include metadata checks and allow customers to disable noisy recommendations.
AI-written summary · about
chapters (12)
- 0:00Empty list information theory
- 1:08Proactive recommendations in OpenShift
- 4:43How proactive recommendations work
- 7:36Why Insights Operator gathers data
- 13:02Empty lists in recommendations
- 19:07Limitations of empty lists
- 20:40Developing recommendations
- 23:15Proactive vs documentation
- 25:39Customizing OpenShift recommendations
- 28:37Updating recommendation rules
- 32:02Preventing future issues
- 35:20Q&A
related talks
25mDave Lindquist, Red Hat and Joe Fitzgerald, Red Hat | Red Hat Summit 2021 Virtual Experience
22mHidden Signals in K8s Clusters
29mGaining Insights from a Black Box System
25mKube Intelligence - A Metric Based Insightful Remediation Recommender
20mGiving your scientific computing environment (SCE) a voice
27mRethinking Observability for Off-the-Shelf Tools
36mDude, You Forgot the Feedback: How Your Open Loop Control...
19mThe Data of Infrastructures