Skip to main content
Question

Chat Errror

  • October 14, 2026
  • 1 reply
  • 28 views
Do Gooder (Customer)
Forum|alt.badge.img+3

We receive a "Failed to load chat version information" error when opening the settings in IFS Cloud 25R2. How can I solve,help me please ?

    1 reply

    Hero (Employee)
    Forum|alt.badge.img+10
    • Jane Perry
    • Hero (Employee)
    • October 14, 2026

    Hi ​@HilalSTS   Do you receive this error for all users?

     

    I did a search on the error and found the following information -

     

    The Copilot chat panel probably can't get version details from the chat backend (the ifsapp-chat container). That usually comes down to setup, permissions or connection to the service. The checks below come from the documented Copilot setup and troubleshooting guides, in the order I'd try them.

    1. Look at the Copilot Error Log first

    Sign in as an admin, search the Navigator for Copilot Error Logs (Solution Manager > Copilot Management > Copilot Error Logs), and find the entry from when the error happened. Note the Log ID, X-Request, error type and error message. The error message usually tells you what to fix next (IFS Copilot Logging).

    By default the log only keeps the last 10 errors. If yours is gone, reproduce the error and check again right away. You can raise the limit in the IFS.ai Copilot Error Log retention policy system parameter (IFS.ai Copilot 25R2 Best practice guide - Internal version).

    2. Check the basic Copilot setup

    From IFS.ai Copilot – set up of Copilot functionality 25R2 SU5 ->:

    • The FNDGPT component is active.

    • The system parameter IFS.ai Copilot Chat is ON.

    • The user has the FNDGPT_RUNTIME permission set.

    • If you're on 25R2 SU5: an admin has to add the projections AllProjections and PermissionSetHandling to FNDGPT_RUNTIME by hand. AllProjections is the name of a single projection, not "every projection." Later service updates include these automatically.

    • Admin settings and publishing prompts also need FNDGPT_ADMIN.

    3. Match the log message to a fix

    These are documented in IFS Copilot Logging:

    Log shows

    Fix

    "NXS_ML_SERVICE or NXS_BASE_URI variables not configured", "Error fetching Nexus access token" or "failed to send request to ML service"

    In the nxs-config ConfigMap, check that uri and mlservice have values and that the deployment maps them to NXS_BASE_URI and NXS_ML_SERVICE. Then restart the ifsapp-chat deployment.

    no-protocol: not-set

    For remote deployments, check nxsUri in custom.yaml (it must include the protocol, e.g. https://), then restart the pods.

    401 Unauthorized

    Check the token, issuer and authentication setup. If the token still can't be verified, escalate to the platform authentication team.

    403 / "Tenant is not authorized to use copilot"

    Check the tenant's status and capabilities. Bring in the platform/release-operations team.

    "You do not have the necessary permissions…"

    Grant the missing permission set or projections (see step 2).

    502 / 503 / 504

    The service isn't ready or can't be reached. Check backend health, then retry.

     

    4. If the log doesn't explain it

    Open the logs of the ifsapp-chat pod (for example in FreeLens), filter for error, and match entries to the X-Request ID. If the chat pod isn't running or keeps restarting, that alone could explain why version information won't load. That last point is my inference, not documented.

     

    If this does not resolve the issue for you, please enter a support case.  Thanks!  Jane

    Terms & ConditionsAccessibility statement