kubectl client. Helm uses your kubeconfig to talk to the cluster, so ensure it points to the intended cluster context.

Make sure
kubectl can talk to your cluster and that your kubeconfig is accessible from your user home directory. Snap-installed Helm (see below) requires a less restrictive sandbox to read kubeconfig by default.- A running Kubernetes cluster (local, cloud, or remote).
kubectlinstalled and configured with the correct kubeconfig.- Network access to download Helm binaries or packages.
Note: The
--classic flag is required for Snap installs to relax confinement so Helm can read files from standard locations (like kubeconfig) on the host.
Install examples (step-by-step)
- Snap (Linux):
- Debian / Ubuntu (APT):
- FreeBSD:
- macOS:
- Windows (PowerShell with Chocolatey):
If you installed Helm via Snap and Helm cannot access your kubeconfig, the Snap confinement may be blocking access. Use the
--classic option on install or place kubeconfig in a location the snap can read. Always verify kubectl can reach the cluster before troubleshooting Helm connectivity issues.- Ensure
kubectlworks first: most Helm connectivity problems are kubeconfig/context related. - Helm 3 removed the server-side Tiller component; everything runs client-side communicating with the Kubernetes API.
- Keep Helm updated via your chosen package manager to receive chart tooling and security fixes.
- Official Helm documentation: https://helm.sh/docs/
- Kubernetes documentation: https://kubernetes.io/docs/
- In a lab environment, install Helm using one of the methods above, add a chart repository (for example Bitnami), and deploy a simple chart (e.g., nginx) to validate end-to-end functionality.