1. Proposed title of this feature request

Add ability to enforce file or file handle limits per container/namespace
2. What is the nature and description of the request?

We have customers who would like the ability to enforce file and file handle limits per container/namespace in a kubernetes/Openshift native way. Today the only way to enforce file limits in Openshift would be node scoped. As an example, by applying a crio.conf.d ulimit via a machineconfig, all container processes would pick up this configuration.

The feature should allow a set of processes (preferably via a kubernetes scope such as pod/namespace) to have a different ulimit from other processes on the node.

3. Why does the customer need this? (List the business requirements here)

Would allow cluster administrators to change the ulimit for 'bad actor' workloads in the cluster that do not follow best practices.

4. List any affected packages or components.

N/a