AI Troubleshooting for Industrial Networks User Guide

PDF

AI Troubleshooting for Industrial Networks User Guide

Common issues and first actions

Want to summarize with AI?

Log in

Find likely causes, the first action to try, and follow-up guidance for sign-in, alerts, thread context, topology, inventory, AI responses, procedure conflicts, and missing evidence.


Use this table to identify a likely cause and the first action to try. Follow the additional guidance under the matching symptom only when the first action does not resolve the issue.

Common issues

Use this table to identify a likely cause and the first action to try. Follow the additional guidance under the matching symptom only when the first action does not resolve the issue.

Symptom

Likely causes

Try this first

If the issue continues

Cannot sign in

User session, credentials, access policy, or authentication issue.

Refresh the page and sign in with your assigned credentials.

Ask an Admin to verify your role or activation status.

No alerts appear

Filters hide alerts, no current alerts exist, syslog forwarding is not configured, or alert processing is delayed.

Open Alerts and clear restrictive filters.

  1. Select All or All Alerts.

  2. Select Unresolved only when you need active alerts.

  3. Ask an Admin to confirm that switches or the site's syslog server are forwarding events to AI Troubleshooting.

Thread context looks wrong

The wrong alert or device was opened, or the thread lacks required context.

Reopen the intended alert thread and verify the device in the report.

  1. Return to Alerts, open the intended alert, and confirm the affected equipment in the report and its embedded visualizations.

  2. If you are an Admin and the device name or location is wrong, check Inventory before continuing.

Topology view is blank or stale

Discovery has not run, inventory is missing, LLDP is disabled, or the data is out of date.

Refresh Topology and search for a known device.

  1. Refresh the view and search for a known device when those controls are available.

  2. Ask an Admin to verify that discovery completed, inventory is current, and the required discovery protocol is enabled.

Inventory search misses a device

The device was not imported, was renamed, has the wrong zone or type, or discovery has not synchronized.

Clear filters and search by IP address, vendor, or model.

  1. Search by IP address, vendor, or model.

  2. Ask an Admin to review the latest discovery result or correct the inventory record according to site procedures.

AI response is incomplete

Required evidence or asset context is unavailable.

Review the available evidence and context.

If the response remains incomplete, treat it as unverified when required evidence is missing. Admins can use a more specific supported prompt and include the equipment name, location or zone, switch, port, and symptom. OT users should use the available preset follow-up actions for the alert.

Recommendation conflicts with site procedure

Local operating, safety, or change-control requirements take precedence.

Follow the approved site procedure.

Follow the approved site procedure.

Do not perform an action that could affect production or safety systems simply because it appears in an AI recommendation.

Record the conflict and escalate it for runbook or configuration review.

Expected data is missing

Evidence is unavailable or a supporting service cannot retrieve the required context.

Capture the alert title, timestamp, affected equipment, thread, and visible error.

Provide the information to an Admin or the next support team.