Lazy Koala is a lightweight framework for root cause analysis in distributed systems. This provides all the tooling needed for RCA from instrumentation to storage and real-time processing of telemetry data using deep learning.
git clone https://github.com/monitoring-projects/lazy-koala
cd lazy-koala/charts
helm install lazy-koala -n lazy-koala --create-namespace .
./charts/lazy-koala - Helm Chart
./control-plane - Kubernetes Operator which binds all the sub-components
./documentation - Thesis written by the author about project
./gazer - eBPF based telemetry extraction agent
./inspector - Simple HTTP proxy to contact kubeAPI from the UI
./scripts - Helper Scripts
./sherlock - Inference agent to calculate anomaly scores
./ui - User dashboard to visualize the system