Agents are enrolled to your tenant. Naming one here issues a token that identifies both the agent and the tenant it belongs to; the agent will only ever receive, and can only ever report against, your own checks.
This has to be reachable from the agent's machine, which is not the same as reachable from your browser. Change it if the agent sits on another network.
This is what runs as root on the target host. It is generated for this agent, and the token below is already baked in.
The token is shown here once and never again. If you lose it, delete the agent and enroll a new one.
Tags are stamped onto every result an agent reports from the moment they are saved; results already recorded keep the tags they were reported with. Deleting an agent revokes its token immediately - it can no longer receive checks or report results, and its past results are kept.
/api/install, authenticated with the agent's own token.agent.env (mode 600) under the running user's own directory — ~/.config/monitoring-agent on Linux — rather than the command line, and passes it to the container with --env-file.monitoring-agent-<name>, restarting unless stopped. No root: Docker has to be installed already and usable by whoever runs the command.Where failure alerts go, for checks with alerting switched on under Alerts on the check itself. Every channel ticked here receives them. Credentials are stored for this tenant only, encrypted. Send test alert uses the fields as they stand, saved or not, so a channel can be proved before anything is written.
Response time (ms)
Recent checks
| Time | Agent | Result | Response | Details |
|---|