Troubleshoot by symptom

Start with what you can observe. A failed request, an offline agent and a healthy check with no incidents require different actions.

The portal page crashes or will not load

If the page shows This page couldn’t be displayed, choose Retry page. Navigation remains available so you can open another area. Unsaved changes in the failed page may need to be entered again. If retry fails, use Reload application to request the application again.

For a failed data request, use the panel’s retry control first. A refresh warning means the displayed data may be older than the current state. An inventory’s loaded time describes when the portal fetched the list; it does not prove that an agent is online or a measurement is recent.

If the problem persists, record the page URL, approximate time and failed request’s status code. Do not include session cookies, authorization headers or tokens in a support request.

An agent is missing or stops reporting

Check Settings → Agents, the status filter and last-seen time. Active is an enrollment status, not proof of a live connection. Check the agent’s endpoint, credential and logs before reinstalling. Follow agent troubleshooting.

For clones or frequently replaced desktops, read the VDI lifecycle guide. Reusing a hostname is different from sharing an enrolled credential.

A test has no measurements

Open the test’s Manage view and check its agent assignment, enabled checks and interval. Confirm that the selected time range includes a scheduled execution and that the agent supports the check. Follow creating tests and reading results.

No data is not a healthy result. Keep the difference visible when presenting evidence to colleagues.

A chart is blank or a query fails

Confirm the time range, filters, source and dimensions. Broaden one constraint at a time. Retry a failed request before changing the query. Use Metrics Explorer and the query reference.

An incident fires repeatedly or will not recover

Compare its transitions with raw measurements and the rule’s recovery criteria. Check delivery history separately from the underlying condition. Acknowledgement records ownership; manual resolution does not fix the service and a continuing breach can open another incident. See incident investigation and health rules.

The estimate is higher than expected

Check whether desktops remain reporting after logout, whether the always-on and temporary cohorts overlap, and whether discovery generates more data than the chosen telemetry profile. Review budget assumptions.

Prepare a useful support request

Include the approximate time and timezone, affected test or agent IDs, what you expected, what happened and the steps already tried. Remove credentials and personal data from logs. A screenshot should include the time range and filters. Use Contact for an enquiry; this does not automatically upload portal evidence.