It would be great to be able to "cache" directories between job executions.
In some cases it helps to greatly speedup job execution.
Similar to Travis: https://docs.travis-ci.com/user/caching/.

Now I achieve it with persistent volume and pipeline function doing explicit (re)store steps.

What can be added to kubernetes-plugin:
-option to manage caches (specifically drop cache for specific job)
-add DSL construct like podTemplate(... cachePaths: ["A", "B/C"])
-default strategy for cache management (shared NFS-backed volume or provisioned PVC per job)