Get started

StormForge collects Kubernetes observability metrics and uses machine learning to generate rightsizing recommendations for CPU and memory resource settings — helping teams reduce cloud spend and improve workload reliability without manual tuning.

What you’ll need

• A Kubernetes cluster (any distribution: EKS, AKS, GKE, minikube, kind, and others)
• A StormForge account — sign up for free at app.stormforge.io/signup
• helm installed locally

In enterprise environments, verify firewall and proxy requirements before installing.

In these docs

• Quickstart — get recommendations in under 10 minutes
• Installation — install StormForge on your clusters
• Configure — set optimization goals, schedules, and apply thresholds at cluster, namespace, or workload level
• Concepts — recommendations, learning periods, the Applier, and how the ML works
• Guides — task-oriented guides for specific scenarios
• Reference — troubleshooting, best practices, and FAQ
• CLI — manage workloads and recommendations from the command line, including an MCP server for AI agents
• Release notes — changelog for StormForge and the StormForge CLI

Not ready to install? The sandbox lets you explore a live environment with view-only access.