Synced from Eclipse Che issue

https://github.com/eclipse/che/issues/20381

Describe the bug

Our namespace user have quota, so we need specify resources (cpu, memory) when we start pods.
In section server, we have added:

  customCheProperties:
     CHE_WORKSPACE_DEFAULT__CPU__LIMIT__CORES: '1'
      CHE_WORKSPACE_DEFAULT__CPU__REQUEST__CORES: '300m'
      CHE_WORKSPACE_SIDECAR_DEFAULT__CPU__LIMIT__CORES: '300m'
      CHE_WORKSPACE_SIDECAR_DEFAULT__CPU__REQUEST__CORES: '50m'

After lauch workspace, it failed because of to pods have no resources specified:

• remote-runtime-injector: no memory and no cpu resources (request and limit)
• che-plugin-artifacts-broker: no cpu resource (request and limit)

Che version

other (please specify in additional context)

Steps to reproduce

1. Deploy Eclpise che 7.35.1 with chectl and add the following options:

spec:
  k8s:
    server:
      customCheProperties:
        CHE_WORKSPACE_DEFAULT__CPU__LIMIT__CORES: '1'
        CHE_WORKSPACE_DEFAULT__CPU__REQUEST__CORES: '300m'
        CHE_WORKSPACE_SIDECAR_DEFAULT__CPU__LIMIT__CORES: '300m'
        CHE_WORKSPACE_SIDECAR_DEFAULT__CPU__REQUEST__CORES: '50m'

1. Create user namespace with quota and without default container quota.
2. Start workspace.

Expected behavior

We expect that is run workspace, ans we expect that the internal pod as juste the required ressource setted like is already the case for ``che-machine-exec{{, theia-ide}}, etc.

Runtime

Kubernetes (vanilla)

Screenshots

No response

Installation method

chectl/latest

Environment

Linux

Release Notes Text

On clusters with CPU quota specified for users namespaces the plugin artifact broker failed to run if the default CPU request is too high. To fix that Che specifies the request and limits for that container.