# Robusta — Smarter Kubernetes Alerts with AI Enrichment

> Robusta enhances Prometheus alerts for Kubernetes by automatically grouping related alerts, enriching them with pod logs, resource graphs, and AI analysis, and routing them to Slack, Teams, or PagerDuty.

## Install

Save as a script file and run:

# Robusta — Smarter Kubernetes Alerts with AI Enrichment

## Quick Use
```bash
# Install Robusta via Helm
helm repo add robusta https://robusta-charts.storage.googleapis.com
helm install robusta robusta/robusta -f generated_values.yaml --set clusterName=my-cluster
# Generate config at https://platform.robusta.dev
```

## Introduction
Robusta sits between Prometheus/AlertManager and your notification channels, enriching every alert with Kubernetes context. When a pod crashes, Robusta automatically attaches logs, resource graphs, and related events to the alert before sending it to Slack, Microsoft Teams, or PagerDuty.

## What Robusta Does
- Intercepts Prometheus alerts and enriches them with pod logs, events, and graphs
- Groups related alerts to reduce notification noise
- Runs automated remediation playbooks for common failure patterns
- Provides AI-powered root cause analysis summaries for each incident
- Offers a web dashboard for browsing enriched alerts and cluster health

## Architecture Overview
Robusta deploys as a Helm chart with two main components: a runner that executes Python-based playbooks in response to alerts or Kubernetes events, and a forwarder that pushes enriched alerts to configured sinks. The runner watches AlertManager webhooks and the Kubernetes API, triggers matching playbooks, and enriches alerts with data gathered from the cluster.

## Self-Hosting & Configuration
- Install via Helm with a generated values file from the Robusta web platform
- Connect to existing Prometheus and AlertManager instances
- Configure notification sinks: Slack, Microsoft Teams, PagerDuty, Opsgenie, or webhooks
- Write custom playbooks in Python for organization-specific automation
- Enable the AI analysis feature to get GPT-powered summaries of alert causes

## Key Features
- Alert enrichment adds logs, graphs, and events to every notification automatically
- Smart grouping correlates related alerts into single actionable incidents
- Built-in playbooks for common issues like CrashLoopBackOff, OOMKilled, and node pressure
- Custom Python playbooks for automated remediation specific to your workloads
- Timeline view shows the sequence of events leading to each incident

## Comparison with Similar Tools
- **AlertManager** — routes and deduplicates alerts; Robusta adds enrichment, AI analysis, and remediation on top
- **PagerDuty/Opsgenie** — incident management platforms; Robusta enriches alerts before they reach these tools
- **Komodor** — commercial Kubernetes troubleshooting; Robusta is open source with a free tier
- **Datadog** — commercial monitoring with built-in alerts; Robusta works with existing Prometheus setups
- **K8sGPT** — AI diagnostics for cluster issues; Robusta focuses on alert enrichment and automated response

## FAQ
**Q: Does Robusta replace Prometheus?**
A: No. Robusta complements Prometheus by enriching its alerts with additional Kubernetes context before routing them.

**Q: Can I write custom automation playbooks?**
A: Yes. Playbooks are Python functions that receive alert or event data and can take any action: restart pods, scale deployments, or call external APIs.

**Q: Does the AI analysis send data externally?**
A: The AI feature sends anonymized alert summaries to generate root cause analysis. It can be disabled if data privacy is a concern.

**Q: What notification channels are supported?**
A: Slack, Microsoft Teams, PagerDuty, Opsgenie, Discord, Google Chat, and generic webhooks.

## Sources
- https://github.com/robusta-dev/robusta
- https://home.robusta.dev

---
Source: https://tokrepo.com/en/workflows/ce6e1b7c-3c0d-11f1-9bc6-00163e2b0d79
Author: Script Depot