// the find
cilium/cilium-cli
CLI to install, manage & troubleshoot Kubernetes clusters running Cilium
The official CLI for Cilium, the eBPF-based Kubernetes networking and security layer. It wraps Helm chart operations into typed subcommands (install, upgrade, status, connectivity test) and adds Cilium-specific diagnostics that kubectl alone can't give you. If you're running Cilium in production, this is the tool you use day-to-day.
The connectivity test suite is genuinely useful — it deploys real test pods and validates actual packet flows using Hubble, so you're testing the data plane, not kubectl describe output. Auto-detection of cluster type (GKE, EKS, kind, k3s, etc.) works well in practice and sets sane Helm defaults you'd otherwise have to look up. The dry-run flags (--dry-run, --dry-run-helm-values) let you audit exactly what gets applied before touching the cluster. ClusterMesh management is a notable differentiator — connecting two clusters and syncing certificates is a multi-step nightmare without this CLI.
It's tightly coupled to Cilium's Helm chart release cadence — the CLI version matrix means you often can't use the latest CLI against an older Cilium install without hitting subtle incompatibilities. No offline mode: everything assumes live API server access, which is annoying in air-gapped or break-glass scenarios. The connectivity test can be slow (several minutes) and gives you noisy flow logs without a clear summary of what actually failed when something goes wrong. Rancher is explicitly unsupported with no timeline, which matters if you're in an enterprise shop standardized on it.