Choose compute / managed infrastructure

Self-hosted Worker Nodes vs managed compute on dFlow Cloud; attach compute to Environments and separate from self-hosting the control plane.

Written By Zoro

Last updated 3 days ago

Services run on Worker Nodes attached to Environments

On dFlow Cloud, you either connect servers you operate (self-hosted nodes) or use managed compute when your plan and region support it. The control plane always lives on dFlow Cloud; this page is only about where workloads execute.

How this differs from self-hosting the control plane

Self-hosting dFlow means running the dashboard and APIs yourself (for example with Docker Compose). That path is documented under Self-Hosting dFlow under Self-Hosting dFlow in the sidebar. Here you stay on app.dflow.sh and only decide which machines back your Environments.

1. Understand the placement model

Skim How dFlow is structured under Core Concepts in the sidebar if you have not already.

  • Application: Groups Environments (for example production vs staging).
  • Environment: Where you attach compute; all Services in that environment run on that pool of Worker Nodes.
  • Worker Node: A Linux host dFlow reaches over SSH (labels in the UI may still say server).

Legacy Projects: If you still use legacy Projects, map them to Applications and Environments using Legacy Projects vs Applications under Core Concepts in the sidebar and plan Migrate from legacy Projects to Applications under Migration and Release Notes in the sidebar when ready.

2. Choose a compute path

Option A: Self-hosted Worker Node (your server or VM)

Use this when you already have a VPS, bare metal, or cloud VM and want full control of the machine.

  1. Prepare a supported Linux host and SSH access; see Add a self-hosted server under Worker Nodes and Compute in the sidebar and Worker Nodes overview under Worker Nodes and Compute in the sidebar.
  2. In the dashboard, start Add server / Add Worker Node (wording may vary).
  3. Complete the connection test and save.
  4. Attach the node to the right Environment: Attach compute to an environment under Environments in the sidebar.

Expected outcome: The Environment lists your node as available for Deployments.

Option B: Managed compute

Use this when dFlow provisions capacity for you (availability depends on plan and region).

  1. Open the managed compute or onboarding flow from the product (exact entry points change with the UI).
  2. Pick region and size options the wizard offers.
  3. Confirm the Environment that should use the new capacity.

Read Managed compute under Worker Nodes and Compute in the sidebar for the detailed concept page as it is expanded.

Expected outcome: Deployments target dFlow-managed capacity without you SSHing into the host for day-to-day use.

3. Validate before you deploy Services

  • At least one Worker Node (self-hosted or managed) is healthy in the UI.
  • The Application’s Environment has compute attached.
  • SSH keys under Security and Team Management in the sidebar and network paths allow the control plane to reach self-hosted nodes when you use them.

If something fails

  • Compute and worker node issues under Troubleshooting in the sidebar
  • Getting started issues under Troubleshooting in the sidebar

Related