Robusta AI

Short Intro: Robusta AI is a Kubernetes observability platform built by Natan Yellin, a former GNOME Linux contributor, founded in 2021 in Tel Aviv. The project ships two products: Robusta Classic, an Apache 2.0 rule-based Prometheus alert enrichment engine, and HolmesGPT, an AI SRE agent that became a CNCF Sandbox project in January 2026 with major contributions from Microsoft. Both can be installed together or independently, and the platform's zero-additional-instrumentation design integrates directly with existing Prometheus, PagerDuty, and Kubernetes tooling without requiring new agents or data pipelines.

Key Capabilities:

• Prometheus alert enrichment with pod logs, graphs, and remediation hints added automatically

• Smart alert grouping into Slack threads to reduce notification spam

• Problem detection without PromQL for OOMKills, failing jobs, and Kubernetes-native events

• Self-healing auto-remediation rules for faster incident resolution

• Change tracking correlating alerts with infrastructure and application changes

• HolmesGPT SRE agent for AI-powered root cause analysis across any stack

• HolmesGPT Operator Mode for 24/7 proactive background monitoring with Slack notifications

• GitHub integration allowing HolmesGPT to open PRs when it identifies fixable issues

• PagerDuty integration for AI-assisted troubleshooting within existing incident workflows

• Skills extensibility system for custom tool integrations via configuration

• MCP integrations for Kubernetes cluster querying, Atlassian/Jira, and custom toolsets

• Kubernetes MCP server exposing cluster data to external AI agents and tools

• CLI-first interface compatible with any OpenAI-compatible LLM including Claude and GPT

• On-premise deployment for teams with data residency requirements

See Robusta AI pricing details →