Using same token to deploy multiple hosts not registering docker

[SKProCH]

Describe the bug
When registering multiple VMs using the same token, only the first one registers as "docker" host.
Other's prints that they detected docker, but actually doesn't add docker hosts:

I've added all this VMs using same token, all have docker installed:

To Reproduce
Steps to reproduce the behavior:

1. Go to 'Settings > Agent'
2. Click on 'Generate token'
3. Copy linux install script
4. Install it on several VMs
5. Only first VM registered as "docker"

Expected behavior
All VMs that have docker should be registered as docker agents

Environment:

• Pulse Version: 5.0.6
• Installation Type: Docker

[pulse-triage]

This is expected behavior. Each Docker agent requires its own unique API token. When you install the unified agent on multiple VMs with the same token:

• Host metrics work (appear as managed hosts)
• Docker metrics are rejected with "API token is already in use by agent"

The agent logs this error, but you might not notice unless checking journalctl -u pulse-agent.

Solution: Generate a unique token for each VM in Settings > Agent before installing.

---

Closed by automated triage. Feel free to reopen if this needs further attention.

[SKProCH]

Yeah, got it. But this is not an intuitive way.

From the first sight everything working fine, docker detected, everything works, but actually didn't.
Probably need to add warning/error to installation, when docker detected, but couldn't be setup due to token reuse?

Also, would be nice, if there is any way to just copypaste same install line and make everything happen automatically. I've have an ~20 lxc containers, and generating the token and copypasting it to every lxc manually is tedious job.
Probably some cmd argument can specify, that installer should generate a separate token for docker installation. Or provide any other way to quickly setup agent on multiple lxc containers (probably some GUI button/toggle in proxmox agent settings, that will install agent on all or nevely added lxcs).

[rcourtman]

Valid points. A few options to consider for your 20-container scenario:

1. Ansible/scripting: Generate unique tokens programmatically via the API (POST /api/agent/tokens) and pass to each installer
2. Host-only mode: If you only need host metrics (not Docker), the same token works fine across multiple hosts

The UX suggestions (warning on token reuse, auto-token generation) are reasonable but out of scope for this issue. Feel free to open a feature request if you'd like these prioritized.

[rcourtman]

Pushed a fix (0a9643c) - when Docker registration fails due to token reuse, the agent now logs a clear error message instead of the generic "Failed to send docker report":

DOCKER REGISTRATION FAILED: This API token is already used by another Docker agent.
Each Docker host requires its own unique token.
Generate a new token in Pulse Settings > Agents and reinstall with the new token.

For bulk deployment to 20+ containers, you can script token generation via the API:

# Generate token
TOKEN=$(curl -s -X POST "https://pulse/api/agent/tokens" \
  -H "Authorization: Bearer $ADMIN_TOKEN" \
  -d '{"name":"lxc-'$HOSTNAME'"}' | jq -r .token)

# Install with generated token
curl -sSL https://pulse/install.sh | bash -s -- --token "$TOKEN"