Running a fleet of client VMs (agency playbook)

Doc-vs-code note: there is no separate "fleet" feature in the UI — this guide describes how to use the existing App Servers list to manage many client VMs under one stack. If you were expecting a dedicated fleet dashboard, it doesn't exist yet; everything below is the same App Servers page described in Bring your own VM, used at scale.

What it's for

If you run an agency and host multiple clients' VMs, you don't need a separate tool per client. Connect (or order) each client's VM as its own App Server inside one project, and get unified monitoring/logs/metrics across all of them from the same console.

How to set it up

1. For each client VM: go to App Servers → Add VM, and either connect their existing server or order a new one on their behalf.
2. Repeat for every VM you manage. They all show up in one App Servers list/table with Name, Status, VM IP, Portainer link, and heartbeat.
3. For any VM with an existing compose stack, use Adopt existing workloads to bring each client's services in as individually monitored applications.
4. Set up Monitoring per app so you have per-client alerts and dashboards you can show as a reliability guarantee.

Gotchas

• There's no per-client grouping, tagging, or sub-project boundary in the App Servers list today — naming convention (e.g. prefixing server names with the client name) is your only organizational tool.
• All servers in a project share the same Members/Roles access — you can't currently give one team member access to only one client's VM within the same project. Use separate projects per client if you need hard isolation (see Members and roles).
• Deleting a VM is a row action on the list — double check you're deleting the right client's server; there's no per-VM confirmation beyond the standard delete confirm.

Not yet supported

• A dedicated multi-tenant "fleet" view, client tagging, or per-VM access scoping within a single project.
• Bulk actions (e.g. re-scan/discover across all VMs at once).