OpenRCA

Root Cause Analysis for Kubernetes

OpenRCA attempts to facilitate root cause inference of issues emerging in Kubernetes clusters. By constructing a real-time cluster topology graph enriched with telemetry data obtained from integrated telemetry sources (Prometheus, Elasticsearch, Falco, Istio, and others), provides operators with a powerful analytical toolkit for dealing with daily application failures, bottlenecks, and misconfigurations. Moreover, it comes with algorithms that leverage the collected graph data to pinpoint sources of complex cluster defects.

Objectives

• Real-time cluster topology visualization
• Integration hub for telemetry data (Prometheus, Elasticsearch, Falco, Istio, and others)
• Post-mortem analysis
• Automated root cause inference
• Diagnostics framework for common applications (databases, load balancers, message queues)
• Site reliability engineering and chaos testing

Installation

Install using Helm chart:

$ helm install ./helm/orca --namespace rca --name orca

Use a 2.x version of Helm. Helm 3 is not supported.

Usage

Dashboard

Port-forward the dashboard for access via the web browser:

$ kubectl -n rca port-forward svc/orca-ui 8080

The dashboard should be available at http://localhost:8080.

Contact

Reach project contributors via these channels:

• Gitter chat room
• Github issues