Red Hat Summit Red Hat SummitSupportConsoleDéveloppeursCommencer un essaiContact

Ce contenu n'est pas disponible dans la langue sélectionnée.

Chapter 1. Troubleshooting OpenShift Lightspeed

Review solutions and workarounds for common installation, configuration, and operational issues encountered with OpenShift Lightspeed.

1.1. 502 Bad Gateway errors in the interface
Copier lien

Wait for service pods to finish starting to avoid 502 Bad Gateway errors.

Wait a few minutes after deploying OpenShift Lightspeed and OpenShift Container Platform before trying the interface again.

1.2. Operator missing from the OperatorHub list
Copier lien

The OperatorHub displays the OpenShift Lightspeed Operator only for supported architectures. Filtering prevents the Operator from appearing on anything other than the x86_64 architecture.

1.3. Reasoning model generates delineator prompt
Copier lien

Reasoning models use tags such as THOUGHT or reasoning to separate inner logic from the final answer.

OpenShift Lightspeed does not control these tags or add them to the output. This feature is part of the model itself.

Procedure

You can turn off these tags using one of the following methods:

  • Add a keyword to your prompt if the model supports it, such as /nothink. Check your model documentation for the specific keyword.
  • Disable the delineator feature in the inference server configuration settings. For more information, see the documentation for the inference server or for the model you are using.

1.4. Troubleshoot API authentication failures
Copier lien

Use the status codes and error details to identify and resolve common authentication failures when connecting to the OpenShift Lightspeed API.

Expand
Status codeDescriptionExample detail

401 Unauthorized

The Authorization header is missing, malformed, or does not use the Bearer scheme.

Unauthorized: No auth header found

403 Forbidden

The token is invalid, expired, or the user lacks RBAC permissions for the /ols-access path.

Forbidden: User does not have access

500 Internal Error

An unexpected error occurred through the Kubernetes TokenReview process.

Forbidden: Unable to Review Token

1.5. Resolving prompt is too long errors
Copier lien

To resolve the Prompt is too long error, adjust the model parameters or reduce the input length for the query to fit within the supported context window.

This error occurs when the total number of tokens (the input query, RAG context, and expected response) exceeds the model context window.

Procedure

  1. Verify that you have set the context window value correctly for your specific model and provider.
  2. Set a lower value for the maximum response tokens parameter to allow more space for the input query and context.
  3. Shorten the query or reduce the size of any attached files.

1.6. Resolving truncated responses
Copier lien

To resolve truncated or incomplete model responses, increase the response token limit or use follow-up prompts to retrieve the remaining content.

This issue occurs when the model reaches its pre-configured response token limit.

Procedure

  1. Verify that the model supports a higher response token limit.
  2. Increase the token limit value in the OpenShift Lightspeed configuration.
  3. If the response is still cut off, type continue as a follow-up query to prompt the model to provide the remaining text.
Note

Set the response token value in reasonable proportion to the context window value. Setting this value too high reserves tokens and might limit the size of your input query.

1.7. Resolving issues with conversation history
Copier lien

To maintain conversation history, optimize the balance between the context window and response tokens. This prevents earlier dialogue from being truncated when the model reaches its context limit.

Procedure

  1. Verify the context window is correctly set for your specific model and provider.
  2. Lower the max response tokens value to increase the remaining space available for conversation history.

1.8. Google Vertex AI configuration resource is rejected
Copier lien

Review common validation errors if the OLSConfig custom resource (CR) is rejected during deployment:

  • Error: googleVertexConfig is required for google_vertex provider

    • Resolution: You must provide the googleVertexConfig object containing both projectID and location when using type: google_vertex.

  • Error: googleVertexConfig may only be set when type is google_vertex

    • Resolution: Remove googleVertexConfig. For the google_vertex_anthropic type, you must use googleVertexAnthropicConfig instead.

  • Error: credentialKey must not be empty or whitespace

    • Resolution: Provide a valid key name string, or omit the field entirely to default to apitoken.
Red Hat logoGithubredditYoutubeTwitter

Apprendre

Essayez, achetez et vendez

Communautés

À propos de Red Hat

Nous proposons des solutions renforcées qui facilitent le travail des entreprises sur plusieurs plates-formes et environnements, du centre de données central à la périphérie du réseau.

Rendre l’open source plus inclusif

Red Hat s'engage à remplacer le langage problématique dans notre code, notre documentation et nos propriétés Web. Pour plus de détails, consultez le Blog Red Hat.

À propos de la documentation Red Hat

Nous aidons les utilisateurs de Red Hat à innover et à atteindre leurs objectifs grâce à nos produits et services avec un contenu auquel ils peuvent faire confiance. Découvrez nos récentes mises à jour.

Legal Notice

Theme

© 2026 Red HatRetour au début