The goal of this RFE is to not have any client impact at all during any planned or kubernetes driving events, some examples of these are;

• Patching underlying nodes
• Platform upgrades
• Autoscaling/HPA
• Pod spec updates which are non-impacting but require deployment rollouts restarts

The end user experience would be;

• 0 API calls dropped
• User via GUI doesn't see any interruptions, or minimal (eg prompted to refresh)
• long running jobs don't report Error when task pods cycle

Raising appropriate RFE's to ensure the customer is tracked and reviewed appropriately.
Impact will be a much better the customer's client experience and the ability for the customer to perform changes knowing there will be no platform impact.