Skip to content

[BUG] baseUrl ignored when specifying it in openai provider block during helm chart install#2398

Description

@jeffoxenberg

馃搵 Prerequisites

  • I have searched the existing issues to avoid creating a duplicate
  • By submitting this issue, you agree to follow our Code of Conduct
  • I am using the latest version of the software
  • I have tried to clear cache/cookies or used incognito mode (if ui-related)
  • I can consistently reproduce this issue

馃幆 Affected Service(s)

Multiple services / System-wide issue

馃殾 Impact/Severity

Minor inconvenience

馃悰 Bug Description

I am running a self-hosted model using vLLM which exposes an OpenAI API compatible endpoint. Creating a ModelConfig after installation works to onboard this model endpoint to kagent but the problem is that during initial installation using the helm chart, the baseUrl for the OpenAI provider is ignored. This makes it not possible to use the default model unless the default ModelConfig is manually updated to add baseUrl after kagent installation.

馃攧 Steps To Reproduce

  1. Deploy a self-hosted model with openai api
  2. Use helm values.yaml similar to below, specifying baseUrl
  3. Observe that the default model config does not use the specified baseUrl by querying the resulting default-model-config. Also observe that default auto-created agents result in 401 error when querying
  4. Manually add baseUrl to default-model-config and observe agents now work correctly
providers:
  default: openAI
  openAI:
    provider: OpenAI
    model: "deepseek-ai/DeepSeek-V4-Flash-0731"
    apiKeySecretRef: kagent-my-provider
    apiKeySecretKey: PROVIDER_API_KEY
    openAI:
      baseUrl: "https://base_url.tld/v1"
    tls:
      caCertSecretRef: llm-certs
      caCertSecretKey: ca.crt
controller:
  substrate:
    enabled: true
    ateApiEndpoint: dns:///api.ate-system.svc:443
    ateApiInsecure: true
substrateWorkerPool:
  create: true
  replicas: 1
  ateomImage: ghcr.io/kagent-dev/substrate/ateom-gvisor:v0.0.6

馃 Expected Behavior

The default-model-config contains the specified baseUrl and default auto-created agents work successfully

馃摫 Actual Behavior

The default-model-config does not contain the specified baseUrl and default auto-created agents fail with 401 error when querying

馃捇 Environment

  • MacOS
  • Kind cluster with k8s v1.34.0
  • Helm chart version 0.9.9

馃敡 CLI Bug Report

No response

馃攳 Additional Context

Manually modifying default-model-config with the baseUrl fixes it but we should be able to specify this at install time

馃搵 Logs

馃摲 Screenshots

No response

馃檵 Are you willing to contribute?

  • I am willing to submit a PR to fix this issue

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    Projects

    Status
    Backlog

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions