Skip to content
Agents tracked: 258 Downloads (7d): 219M up 9.4% GitHub stars: 5.5M VS Code installs: 148M Releases (7d): 282 Agent pull requests (last week): 940K Updated Oct 6, 2026

AI SRE and DevOps agents

Agents that start investigating the moment an alert fires: they query telemetry, check recent changes and propose a root cause.

By GitHub stars, the leaders are K8sGPT (8,218), HolmesGPT (3,511). As of Oct 6, 2026.

# Agent Pulse Downloads 7d 7d 30d VS Code installs Stars Latest release Price Last 90 days
188
K8sGPTK8sGPT (CNCF Sandbox)
17 — — — — 8,218+2/day 0.4.3922d ago Free (OSS) —
203
HolmesGPTRobusta (CNCF Sandbox)
14 2,330 up 24.5% down 26.0% — 3,511+5/day 0.42.020d ago Free (OSS)
218
Resolve AIResolve AI
4 — — — — — — Usage-based —
225
ClericCleric
0 — — — — — — Free + $700/mo —
234
TraversalTraversal
0 — — — — — — Custom —
236
— — — — — — — Usage-based —
248
— — — — — — — $25/mo —

Ranks are overall positions by Pulse Score. Click a column to sort by what matters to you. Methodology.

How an AI SRE works

When an alert fires or an incident is declared, the agent forms hypotheses and tests them against logs, metrics, traces, deploys and past incidents, then posts its findings with evidence. Some also verify new deploys and draft fix pull requests.

Choosing one

  • Data access. Agents inside an observability platform (Datadog) see that platform's data; independent ones connect to several sources.
  • Permissions. Start read-only; let the agent suggest actions before it takes any.
  • Open source. HolmesGPT and K8sGPT are CNCF projects you can run in your own cluster with your own model.

Head-to-head comparisons

Related guides

Questions about AI SRE and DevOps agents

What is an AI SRE?

An AI SRE is an agent that investigates production alerts and incidents: it queries telemetry, correlates recent changes and reports a likely root cause with evidence, so on-call engineers start from a diagnosis.

Is there an open-source AI SRE agent?

Yes. HolmesGPT and K8sGPT are open-source CNCF Sandbox projects that diagnose Kubernetes and production issues with a model of your choice.