-
Feature Request
-
Resolution: Unresolved
-
Undefined
-
None
-
None
-
None
-
False
-
None
-
False
-
Not Selected
-
-
-
1. Proposed title of this feature request :
Enforce Minimum Node Group Size in OpenShift Cluster Autoscaler via --enforce-node-group-min-size Flag
2. What is the nature and description of the request?
The request proposes adding a configurable --enforce-node-group-min-size flag to the Cluster Autoscaler (CA) in OpenShift. This flag would ensure that CA maintains the minimum node count in each node group (e.g., MachineSet), even when there are no unschedulable pods. Currently, CA scales nodes up only when unschedulable pods are present, potentially allowing MachineSets to fall below their configured minimum, which can delay resource availability during workload spikes.
3. Why does the customer need this? (List the business requirements here)
- When demand increases unexpectedly, having a minimum number of nodes already available reduces the time needed to scale up and meet the load, resulting in better performance and responsiveness.
- Reduced Manual Intervention: With this feature, administrators would not need to monitor and manually intervene to ensure minimum node levels, saving time and reducing operational overhead, especially in production clusters.
4. List any affected packages or components.
- Cluster Autoscaler: Specifically, the Kubernetes Cluster Autoscaler implementation within OpenShift.
- MachineSet Configuration: Any MachineSets configured to use Cluster Autoscaler in OpenShift