Uninstall or move a server

Remove the agent from the server

Run the installer with the uninstall flag as root:

curl -sSL https://api.getopservo.com/install.sh | sudo env OPSERVO_UNINSTALL=1 bash

It stops and disables infrapilot-agent.service, deletes the unit file, removes /usr/local/bin/infrapilot-agent, and deletes /etc/infrapilot (the config with the agent’s credentials, plus any mysql.cnf or pg.conf you placed there) and /var/lib/infrapilot (the local offline buffer). It prints Opservo agent removed from this server.

It leaves alone anything it did not create: a sudoers entry you added for remote actions, SSH keys granted through Opservo, and the intel-gpu-tools package if the installer added it. Remove those by hand if you no longer want them.

Remove the server in the dashboard

On the server page open the menu and choose Remove server. The confirmation says exactly what happens: the server and all of its history (snapshots, findings, logs) are deleted from Opservo, and this cannot be undone. The same dialog shows the uninstall command above with a Copy button so you can clean the host at the same time.

If you remove the server in the dashboard but leave the agent installed, the agent keeps running and its heartbeats are simply rejected; it never reappears. Uninstall it to stop the noise.

Chat conversations that were pinned to the server are kept; they just lose the server link. The audit log entries for the removal are kept too.

Archive instead of delete

Archiving takes a server off the fleet grid without deleting its history. An archived server does not count toward your plan’s server limit or per-server billing. You will see archived servers in two situations:

• •An Auto Scaling instance that was enrolled with a fleet key and has been offline past the archive threshold is archived automatically. See Auto Scaling.
• •A server that was moved to another workspace is archived in the workspace it left, and that workspace gets an alert saying so.

Archived servers appear at the bottom of the fleet grid under Archived servers, each with a Restore button. A server also un-archives itself the moment it sends a heartbeat again, so a false positive heals on its own.

What happens to history

Independently of these actions, raw metric snapshots are retained for 30 days.

Re-enrol a server with a new identity

Normally you never need this: re-running the installer reuses the credentials in /etc/infrapilot/config, and moving between workspaces is handled by the Add server dialog. Use OPSERVO_FORCE=1 only when you want the server to register as a new entry, for example after restoring a disk image onto different hardware or after removing the old entry in the dashboard.

1. 1.In the dashboard, click Add a new server and generate a fresh command. The token is required; registration without one is refused.
2. 2.Add OPSERVO_FORCE=1 to the env part of the command and run it on the server:

curl -sSL https://api.getopservo.com/install.sh | sudo env OPSERVO_FORCE=1 OPSERVO_API_URL=https://api.getopservo.com OPSERVO_TOKEN=<token> bash

The installer skips the saved credentials, registers a new agent ID and secret, and rewrites /etc/infrapilot/config. OPSERVO_ENABLE_PROBES=true and OPSERVO_ENABLE_SHELL=true settings are carried over; any other line you added by hand (for example OPSERVO_ENABLE_ACTIONS=false to hard-disable remote actions on the host) has to be added again. The new entry starts with remote actions off in the dashboard, like any new server. The previous entry goes offline and keeps its history until you remove it.

Kubernetes clusters

Removing a cluster from the dashboard deletes its history and findings here, but the collector pod keeps running until you remove it: kubectl delete namespace opservo, plus the opservo-k8s-agent ClusterRole and ClusterRoleBinding. See Kubernetes.