1. Proposed title of this feature request:
Multiple Custom Profiles for Node Descheduler
2. Who is the customer behind the request?
Meijer
3. What is the nature and description of the request?
We would like to create different profiles for different application teams and different application personalities.
4. Why does the customer need this? (List the business requirements here)
We plan to use this to control/balance VM and container soft anti-affinity on our baremetal compact (3 node) clusters at the edge. We use soft anti-affinity because being at the edge there are not a lot of nodes for workloads to move to. Ideally they are on different nodes but worst case we still want the app to come up. After the disruption event is resolved though we would like the workloads to automatically go back to respecting their anti-affinity rules.
5. How would the customer like to achieve this? (List the functional requirements here)
We would like to use multiple custom profiles with the Node descheduler or multiple descheduler CRDs, I am not opinionated either way but multiple custom profiles seemed like the more achievable option.
6. For each functional requirement listed, specify how Red Hat and the customer can test to confirm the requirement is successfully implemented.
We would setup different policies for different applications and observe during the descheduler intervals if the different profiles are respected and reflected.