Quickstart

Supernode is an opinionated platform for Cardano Stake Pool Operators who need a repeatable way to bootstrap infrastructure, install supported workloads, manage secrets, and monitor production behavior from a single operating model.

The preferred way to operate Supernode is to load the repository skills into your agent of choice and ask for the workflow you want.

Typical prompts include:

• “I would like to bootstrap a new Supernode”
• “I would like to bootstrap a new pool from scratch”
• “I would like to migrate my Cardano Preview pool into the Supernode”
• “I would like to deploy Dolos on my Supernode”

The manual guides in these docs remain the canonical reference, but the preferred operator experience is goal-oriented management through the skills shipped in skills/.

Start with Skills if you want the agent-first workflow.

This quickstart follows the canonical Supernode path:

1. Bootstrap a cluster and install the Supernode control-plane

   If you are using an agent, load the Supernode skills first and ask for a bootstrap workflow. The manual commands below explain the underlying reference path.

   kind

   Use kind when you want the fastest path to evaluate Supernode locally.

   cd bootstrap
   ./bootstrap.sh \
     --provider kind \
     --version 0.2.1

   aws

   Use AWS when you want a managed Kubernetes environment for a real deployment.

   cd bootstrap
   ./bootstrap.sh \
     --provider aws \
     --version 0.2.1

   gcloud

   Use Google Cloud when you want to run Supernode on GKE. Autopilot is the default mode.

   cd bootstrap
   ./bootstrap.sh \
     --provider gcloud \
     --version 0.2.1

   This installs the required control-plane, including Prometheus, Grafana, Vault, and Vault Secrets Operator.

   For the full provider-specific flows, see Bootstrap.

2. Complete the control-plane post-install steps

   After bootstrap, finish the Vault setup before expecting producer workloads or Vault-backed secrets to work correctly.

   kubectl -n control-plane get pods
   kubectl -n control-plane exec -it control-plane-vault-0 -- vault status

   For standalone or HA Vault modes, initialize Vault once and then run:

   cd extensions/control-plane
   VAULT_TOKEN=<vault-admin-token> ./scripts/post_install.sh

   For the full sequence, see Control Plane.

3. Install a workload

   The canonical Supernode model is to install supported workloads on top of the control-plane. For example, a Cardano relay:

   helm install preview-relay ./extensions/cardano-node \
     --namespace preview-relay \
     --create-namespace \
     --set displayName="Preview Relay" \
     --set node.network=preview \
     --set node.networkMagic=2 \
     --set extraLabels.supernode/status=ready

   You can also install workloads through the dashboard once it is running in-cluster.

   For an agent-guided install, use the relevant skill instead of manually stitching the workflow together. For example:

   • cardano-relay-setup.md
   • cardano-stake-pool-from-scratch.md
   • dolos-supernode-deployment.md

   For canonical workload patterns, see Extensions.

4. Open the dashboard and monitor the workload

   The Supernode dashboard is the operator entrypoint for installed workloads. It discovers OCI-packaged extensions, installs them into namespaces, shows workload health, and links to Grafana dashboards.

   Start by checking that the dashboard and monitoring stack are up:

   kubectl -n control-plane get pods
   kubectl -n control-plane port-forward service/grafana 3000:3000

   Then use the dashboard to inspect health, logs, and workload-specific telemetry.

   For the full operator workflow, see Dashboard and Monitoring.