[BUG] Configuring ephemeral-storage request/limits for Flyte tasks

[rstanevich]

Describe the bug
A Python task has an option to define storage_request and storage_limit at the arguments: Source
But these values don't appear in the requests/limits parameters of k8s Pod.

Expected behavior
An option to specify ephemeral-storage requests/limits for the Pod running in Kubernetes.
https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/#local-ephemeral-storage

Flyte component

• Overall
• Flyte Setup and Installation scripts
• Flyte Documentation
• Flyte communication (slack/email etc)
• FlytePropeller
• FlyteIDL (Flyte specification language)
• Flytekit (Python SDK)
• FlyteAdmin (Control Plane service)
• FlytePlugins
• DataCatalog
• FlyteStdlib (common libraries)
• FlyteConsole (UI)
• Other

To Reproduce
Steps to reproduce the behavior:

1. Create a python task with argument storage_request:

@python_task(memory_request='2000M', cpu_request='500m', storage_request='1Gi')
    ...

2. Run this in Flyte.
3. We see that FlyteWorkflow CRD instance has these requests parameters in the tasks section:

...
tasks:
  '<< task unique name >> ':
    container:
      resources:
        requests:
          - name: CPU
            value: 500m
          - name: MEMORY
            value: 2000M
         - name: STORAGE
            value: 1Gi
...

4. But nodes section in the FlyteWorkflow instance and created Pod specification don't have ephemeral-storage requests:

resources:
    requests:
        cpu: 500m
        memory: 2000M

Environment
Flyte component

• Sandbox (local or on one machine)
• Cloud hosted
  • AWS
  • GCP
  • Azure
• Baremetal
• Other

Additional context
Discussed with @kumare3 in the Slack. He suggested to file an issue for storage support.
The use case example: a task in which we expect a huge file with geographic data which is very problematic to split.

[kumare3]

Thank you @rstanevich, we will track it for the next release