Documentation menu

Agent Quickstart

The fastest path to a connected cluster, for evaluation installs.

This is the fast path for evaluation installs on a test cluster. For production rollouts — full requirements, network, security, validation, and troubleshooting — see the Installation Guide.

Get the DataAgent agent installed in your Kubernetes cluster and connected to the dashboard in about 5 minutes.

TL;DR

  • In the DataAgent dashboard, open Clusters → Connect, copy the install command, run it.
  • Watch your cluster appear once the agent sends its first heartbeat.
  • The chart brings its own Prometheus, so there is nothing else to install.

You’ll need

  • A Kubernetes cluster (v1.24+) — EKS, GKE, AKS, or anything conformant, on amd64 or arm64 nodes.
  • kubectl configured for the cluster.
  • helm v3.14 or newer.
  • Outbound access from the cluster to tunnel.data-agent.co:443, ghcr.io:443, and the public registries the bundled monitoring pulls from (quay.io, registry.k8s.io, docker.io) — see Prerequisites → Network.

1. Install the agent

In the DataAgent dashboard, open Clusters → Connect. The flow generates a registration token and shows the exact helm install command pre-filled with your token, cluster name, and chart version — copy it and run it on your cluster.

The command looks like this — the dashboard fills in your real values, including the image pull password (you do not create a GitHub token). Paste the dashboard’s command rather than this reference snippet to avoid ImagePullBackOff on first start. The dashboard’s command does not include --wait --timeout 10m; we recommend adding it, so Helm waits until every pod is ready and reports a failed install as an error:

helm upgrade --install dataagent \
  oci://ghcr.io/data-agent-co/dataagent-chart \
  --version <CHART_VERSION> \
  --namespace dataagent --create-namespace \
  --set 'imageCredentials.password=<PULL_PASSWORD>' \
  --set 'config.authToken=<TOKEN_FROM_DASHBOARD>' \
  --set 'config.clusterName=<YOUR_CLUSTER_NAME>' \
  --wait --timeout 10m

2. Verify

kubectl -n dataagent get pods

The agent pod shows 5/5 Running, alongside three monitoring pods: prometheus-server, kube-state-metrics, and one node-exporter per node. Your cluster then appears in the dashboard.

3. Prometheus

Nothing to do — the chart installs its own Prometheus, kube-state-metrics, and node-exporter, and the agent uses it automatically. Shortly after the agent connects, the Prometheus integration card on the cluster detail page goes green.

The bundled Prometheus is the only one the agent integrates with. If you already run your own monitoring stack for your application metrics, keep it — it is optional, and the agent does not connect to it.

What’s next

Cookie Settings