> ## Documentation Index
> Fetch the complete documentation index at: https://docs.odigos.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Switch from Single Node to Multi Node

> Move an existing Agent-only ECS deployment to Agent + Proxy so the cluster appears in Odigos Central as one platform — without reinstalling the agents.

<Info> The Odigos VM Agent is available as part of Odigos Pro. For access and licensing details,
[contact us](https://odigos.io/).</Info>

No reinstall. The agents already running from
[Install: Single Node](/vmagent/ecs/installation-agent-only) speak the same
protocol to the proxy that they speak to Central. You add the proxy, re-point
the agents at it, and roll the daemon.

After the switch, Central shows **one** platform named after the ECS cluster,
replacing the per-node platforms.

<Info>
  You need the **`odigos-ecs-proxy`** image in addition to the agent image you
  already run. Mirror it the same way as in
  [System Requirements](/vmagent/ecs/requirements#license-and-images).
</Info>

## Switch

<Steps>
  <Step title="Deploy the proxy">
    Follow [Install: Multi Node](/vmagent/ecs/installation-with-proxy) steps
    3–5 (stable address, proxy task definition, proxy service). Prefer the
    internal NLB; a pinned instance is test/dev only.

    Do **not** create a second agent service — keep the daemon you already have.
  </Step>

  <Step title="Re-point the agents">
    In the agent task definition, change `ODIGOS_DEFAULT_CONFIG`'s endpoint to
    the proxy's `host:4321` **and** add `ODIGOS_ECS_BEHIND_PROXY=true`. Register
    the new revision.

    The endpoint alone cannot tell a proxy from Central. Without
    `ODIGOS_ECS_BEHIND_PROXY=true`, every agent still registers under the
    cluster name and the coverage check reports all instances as uncovered. See
    [platform naming](/vmagent/ecs/configuration#platform-naming-per-mode).
  </Step>

  <Step title="Roll the daemon">
    ```bash theme={null}
    aws ecs update-service --cluster "$CLUSTER" --service odigos-ecs-agent \
      --task-definition odigos-ecs-agent --force-new-deployment
    ```
  </Step>

  <Step title="Verify">
    Use [Install: Multi Node, Step 8](/vmagent/ecs/installation-with-proxy#step-8-verify):
    agents connected to the proxy, `/healthz` lists every node, and Central
    shows one platform for the cluster.
  </Step>
</Steps>

Values already persisted in each node's `config.yaml`
[take precedence](/vmagent/ecs/configuration#configuration-precedence-and-persistence)
over `ODIGOS_DEFAULT_CONFIG`. After the first boot behind the proxy, change the
endpoint through Central or `odictl` if a node still points at Central.

## Switch back to single node

The same flow in reverse:

1. Point `ODIGOS_DEFAULT_CONFIG`'s endpoint back at Central.
2. Remove `ODIGOS_ECS_BEHIND_PROXY` (or set it to `false`).
3. Register a new revision and `update-service --force-new-deployment`.
4. Delete the proxy service.

Each node then appears in Central as its own platform again.

## Next steps

<CardGroup cols={2}>
  <Card title="Install: Multi Node" icon="sitemap" href="/vmagent/ecs/installation-with-proxy">
    Full Agent + Proxy install if you are starting from scratch.
  </Card>

  <Card title="Configuration Reference" icon="sliders" href="/vmagent/ecs/configuration">
    `ODIGOS_ECS_BEHIND_PROXY` and the proxy environment variables.
  </Card>
</CardGroup>