[Medium/Large Impact Change]: [Discussion] Implementation of changes in Configuration for determining what resources applications need

Description

We need to define:

• What resources each application needs, this includes memory, cores, disk, NIC, etc.
• How this should be defined in configuration? In first instance, we will want still to be tied into using the same host (so we do not attempt to do too much at once), and we can revisit this with Custom Resource Definitions (CRDs) in k8s.
• How this can be mapped into the BootRequest objects.

Related issues and PRs

N/A yet!

Impact radius

Large

Impact on developers and users

The payload associated with a BootRequest will need to include additional metadata including e.g. pod mounts, CPU requirement, etc.

Useful coordination

This will initially be a discussion on how this should be implemented, and then will scale out to the WGs to implement the requested changes

Testing suggestions and/or changes needed in testing

When running the k8s PM with HW configs, they should work

Anything else?

Notes brief, they need expanding

Start by reviewing the BootRequest objects and the Kubernetes process manager's hardware configuration path. Define how memory, cores, disk, NICs, and pod mounts are represented in configuration and mapped into BootRequest metadata. Done means the design is agreed with the working groups and Kubernetes PM runs with hardware configurations.