FAQ
The questions that come up most once you’re actually running clusters: creating and upgrading them, deploying to them, and working out what’s backed up, billed, or on the roadmap.
How do I create a new cluster?
Section titled “How do I create a new cluster?”Open the console, go to Clusters, select Create Cluster, choose the Kubernetes version and resource limits, and submit. Use the cluster status to track progress until it reaches Running.
How do I deploy apps?
Section titled “How do I deploy apps?”Use GitOps with Argo CD. Keep your manifests in Git, create an Argo CD Application, and validate rollout with logs, metrics, and dashboards.
How do I access my cluster from the CLI?
Section titled “How do I access my cluster from the CLI?”From the cluster view, use the CLI access flow and choose:
- OIDC kubeconfig for interactive terminal use (requires
kubelogin, 8-hour access token). - Token kubeconfig for automation (expiry selectable up to 365 days).
See Cluster Access for full setup details.
Can I upgrade a cluster?
Section titled “Can I upgrade a cluster?”Yes. Upgrades are initiated from the console. Plan for a maintenance window and validate workload compatibility before rolling the version forward.
Can I delete a cluster?
Section titled “Can I delete a cluster?”Yes, but deletion is permanent. Confirm backups, data handling, and workload migration first, then delete the cluster from the console.
Are my clusters backed up?
Section titled “Are my clusters backed up?”Not individually, during the alpha. Kupe snapshots the platform’s control-plane state (host etcd, every 12 hours, off-site), but each tenant cluster’s own Kubernetes state is not backed up as a separate, restorable unit yet. If a cluster’s control-plane storage were lost, its objects would have to be re-applied from your source of truth, so keep your manifests, Helm values or GitOps repo authoritative, and treat the cluster as reproducible from git. Application data (databases, volumes) is always your responsibility to back up. Per-cluster backup and restore is on the roadmap for general availability.
How does billing work?
Section titled “How does billing work?”Billing combines:
- A monthly plan base fee.
- Usage-based charges for allocated CPU, memory, and storage.
- Automatic credit application before payment.
See Billing for details.
Who handles payment processing?
Section titled “Who handles payment processing?”Kupe Cloud uses Paddle for payment processing.
Do you support custom domains and TLS?
Section titled “Do you support custom domains and TLS?”Yes. Use HTTPRoute hostnames for routing, and use the custom-domain flow when you
want to serve traffic on your own domain. Kupe handles certificate issuance and HTTPS
for the supported routing paths.
Can I use Terraform or API today?
Section titled “Can I use Terraform or API today?”Yes. Kupe Cloud supports both a Terraform provider and a public API, alongside the console. Use the console for interactive workflows, Terraform for infrastructure as code, and the API for direct automation and custom integrations.
How do secrets work?
Section titled “How do secrets work?”Secrets are stored in a dedicated vault that Kupe manages for you. You define the secret once, then sync it to one or more cluster and namespace targets.
Deleting a secret removes it from the vault, but already-synced Kubernetes Secret objects are not removed automatically. See Secrets.
Where do I view logs and metrics?
Section titled “Where do I view logs and metrics?”Use Grafana for both. Logs are available in Grafana Explore through Loki, and metrics are available through dashboards and PromQL queries.
See Observability.
Where do I troubleshoot issues first?
Section titled “Where do I troubleshoot issues first?”Start with:
- Cluster and workload state in console/Headlamp.
- Logs, metrics, and dashboards in Grafana.
- Routing and domain configuration if the issue is traffic-related.
- The most relevant guide for the workflow you are working on.