Platform overview
Understand the ClusterPilot operating model before you install or enable a lifecycle capability.
Exact release scope. This page documents the reviewed v0.0.1 source at fcc5871. Check release status before enabling a gated capability.
What ClusterPilot is
ClusterPilot is a self-hosted control plane for infrastructure and Kubernetes lifecycle operations. It turns an operator request into a validated plan, applies authorization and policy gates, delivers ordered work to a lifecycle agent, and preserves the resulting evidence.
The platform is designed for customer-controlled environments. The API, operational data, credentials, agents, registry integration, and target clusters remain inside the environment you operate. ClusterPilot does not require a shared SaaS control plane.
What it is not
| ClusterPilot is | ClusterPilot is not |
|---|---|
| A control plane for explicit lifecycle operations | A managed Kubernetes service |
| An evidence-producing execution system | An autonomous operator that changes clusters without approval |
| Self-hosted software with customer-managed dependencies | A claim that installation alone creates compliance |
| A versioned, feature-gated release line | A blanket statement that every coded endpoint is production ready |
Core operating model
| Concept | Purpose | Operator question |
|---|---|---|
| Intent | The requested topology or operation. | What do I want to change? |
| Plan | A stable preview of resources, commands, dependencies, and risk. | What will happen? |
| Guardrail | Authorization, policy, quota, compatibility, and confirmation checks. | Is it allowed and safe enough to start? |
| OperationRun | Durable execution state with ordered steps and recovery semantics. | What is happening now? |
| Evidence | Events, logs, artifacts, checksums, audit records, and correlation IDs. | How can I prove and diagnose the outcome? |
Capability map
Infrastructure & cluster lifecycle
Provider accounts, networks, machines, node preparation, kubeadm bootstrap, scale, upgrade, replacement, and deletion.
Kubernetes operations
Cluster inventory and selected read paths. High-risk mutations remain contract-only or disabled by default.
Add-ons & validation
Curated and custom add-on planning, artifact locks, health gates, dashboard posture, and Sonobuoy evidence.
Self-hosted distribution
Compose and Helm surfaces, digest-pinned images, signed metadata, external PostgreSQL, durable artifacts, and release verification.
Recommended evaluation path
- 01
Read the release status
Confirm that the required capability and provider are inside your approved evaluation scope.
- 02
Choose a deployment profile
Match single-node, split, or HA-ready placement to your availability and operational requirements.
- 03
Complete preflight without mutation
Validate release artifacts, dependencies, storage, trust, time, and configuration before installation.
- 04
Use a disposable first environment
Prove provider, cluster, add-on, conformance, backup, recovery, and cleanup paths before production approval.