Description:

The rhods-dashboard deployment, which is a required component of OpenShift AI and managed by the OpenShift AI Operator upon the creation of a DataScienceCluster (DSC) Custom Resource, currently uses hardcoded CPU and memory resource requests and limits.

These hardcoded values are applied to both containers within the pod (rhods-dashboard and oauth-proxy):

    name: rhods-dashboard
    resources:
      limits:
        cpu: "1"
        memory: 2Gi
      requests:
        cpu: 500m
        memory: 1Gi
...
...
    name: oauth-proxy
    resources:
      limits:
        cpu: "1"
        memory: 2Gi
      requests:
        cpu: 500m
        memory: 1Gi

Problem Statement:

The current fixed configuration leads to two critical scenarios where the dashboard cannot be deployed or perform optimally:

1. Deployment Failure in Constrained/Small Clusters: The cumulative resource requests (2Gi memory, 1 CPU total) can often exceed the capacity of small-scale worker nodes. This results in the rhods-dashboard pods remaining permanently in a Pending state due to scheduling constraints, blocking the use of the product.

1. Performance and Stability Issues in Large/High-Traffic Clusters: In environments with high user concurrency, the default limits (1 CPU, 2Gi memory) may be too restrictive, leading to frequent CPU throttling or OOMKilled events for the dashboard process. Currently, administrators have no mechanism to increase these resource limits to improve stability.

Requested Enhancement:

Introduce optional fields within the DataScienceCluster (DSC) CRD specification to allow users to override the default resource requests and limits for the rhods-dashboard and oauth-proxy containers.

The configuration should be integrated into the DSC CR, for example, under the spec.components.dashboard section.

apiVersion: datasciencecluster.opendatahub.io/v1
kind: DataScienceCluster
metadata:
  name: example-dsc
spec:
  components:
    dashboard:
      managementState: Managed
      # --- NEW OPTIONAL CONFIGURATION BLOCK ---
      resources: 
        rhodsDashboardContainer:
          limits:
            cpu: <string> 
            memory: <string>
          requests:
            cpu: <string>
            memory: <string>
        oauthProxyContainer:
          limits:
            cpu: <string>
            memory: <string> 
          requests:
            cpu: <string>
            memory: <string>

Acceptance Criteria:

1. The OpenShift AI Operator must successfully apply custom requests and limits specified in the DSC CR to the respective containers in the resulting rhods-dashboard deployment.

1. If the proposed resources configuration block is omitted from the DSC CR, the Operator must default to the current hardcoded values to maintain backward compatibility.

Business Justification

This feature is necessary for operability and successful adoption across the full spectrum of customer environments:

• It enables successful installation and deployment on smaller, resource-constrained clusters by allowing users to lower resource requests.

• It ensures the stability and performance of the OpenShift AI dashboard in large, production environments by allowing administrators to tune resource limits to handle high load.

• It improves the overall administrative experience by allowing for Kubernetes best practices regarding resource management and fine-tuning.