Robusta: An Open‑Source Python Platform for Kubernetes Troubleshooting and Automated Incident Response

Robusta is an open‑source Python‑based platform that sits on top of a monitoring stack (Prometheus, Elasticsearch, etc.) to automatically diagnose and remediate Kubernetes alerts.

The project consists of three core open‑source components: an automation engine, built‑in automations for enriching and fixing common alerts, and additional manual troubleshooting tools.

Optional add‑ons include a bundled Prometheus stack, a web UI for viewing alerts, changes and events across clusters, and a CLI.

The automation engine runs as two Kubernetes Deployments (robusta-forwarder and robusta-runner). The forwarder watches the API server and forwards events to the runner, which executes playbooks composed of Triggers, Actions and Sinks (e.g., sending logs to Slack).

Example playbook syntax is shown, and the engine can be installed via Helm, with a generated generated_values.yaml that configures destinations such as Slack.

Installation steps include installing the robusta‑cli (Python ≥ 3.7), generating the Helm values file, adding the Helm repo, and running helm install with optional flags for small clusters.

Robusta monitors a set of Prometheus alerts (CPU throttling, OOM kills, node readiness, etc.) and other errors detected from the API server (CrashLoopBackOff, ImagePullBackOff, Node NotReady), forwarding warnings and events to the UI.

A test deployment demonstrates automatic Slack notifications when a pod crashes, and a custom playbook example shows how to send a Slack message on Deployment updates.

For more advanced usage, users can create custom playbooks and refer to the official documentation at https://docs.robusta.dev.