Troubleshooting

Start with the agent service and its recent logs on the affected server:

sudo systemctl status vpsagent --no-pager
sudo journalctl -u vpsagent -n 50 --no-pager

Check the message before reinstalling. Keep setup tokens, credentials, and webhook URLs out of screenshots or support messages.

The server is missing or reporting delayed

Confirm that the installer finished and the agent service is running. Check outbound HTTPS access to vpsmon.cloud and GitHub. A new installation also needs a valid, unused setup token from your workspace.

Cloud's missing-telemetry timeout uses fresh resource samples. A reporting delay does not confirm that the host is down. A control connection that reports the agent version does not count as a resource heartbeat. If uploads stopped because your workspace's Cloud access is paused, check the account status shown above the dashboard.

Keep the server clock synchronized. Cloud rejects old or repeated collection timestamps even if an agent sends them with a new upload ID. If the service runs but samples remain stale, update the agent and inspect its logs for collection errors.

The installer cannot finish

If it reports an existing installation, run sudo /opt/vpsagent/update.sh. If the earlier install stopped partway through, inspect the service and files before removing that partial installation.

If the token is invalid, expired, or already used, open Add server again for a new token. Paste it only when the installer asks.

If release files are unavailable, check the agent's GitHub releases and outbound network access. Do not repeatedly run the installer over an existing installation.

Disk values differ from the local dashboard

Check the writable disks in Cloud's server details. Cloud alerts use the fullest monitored writable disk. Read-only Snap and ISO images are excluded.

If a real writable disk is missing, update the agent. Historical points collected before a fix remain in the chart until they age out; a new sample does not rewrite old history.

Live processes are empty

Leave server details open for a heartbeat. The agent needs a recent version that supports live process requests. Keep the browser tab visible while waiting.

If resource values arrive but processes do not, update the agent and check its logs. Process collection can fail without stopping normal resource uploads.

An email does not arrive

For incident emails, open Alert settings and check that Email alerts is enabled and the recipient is verified. A different recipient stays pending until its confirmation link is used.

Look at Alert history. Sent means the provider accepted the email; check the recipient's spam folder too. Pending may be waiting for a retry. Failed shows the latest delivery error. Short CPU spikes may not have opened an incident because CPU and memory use averages and confirmation periods.

For account verification or password reset, check the address entered on the sign-in page. Repeated requests are limited by a cooldown and daily quotas. If you still cannot receive a link, contact support.

A webhook fails

Check the webhook setup guide. The receiver must use public HTTPS on port 443 and return a successful 2xx response. Redirects and private network addresses are rejected.

A Slack URL belongs to the channel selected when you created it. Check that the Slack app and channel still exist. For a generic receiver, verify the signature against the raw body and the current signing secret. Changing the destination URL or format rotates that secret.

Contact support

Email support@vpsmon.cloud with the server's display name, agent version, the error message, and roughly when it happened. Include whether all servers are affected or only one. Remove passwords and tokens from any logs you attach.