Skip to content

Scaling Lab Fix

Compare
Choose a tag to compare
@RichardExley RichardExley released this 29 Nov 18:40
· 480 commits to main since this release
6df4dbc

Order-helidon was scaling on a single k8s node. Add a topologySpreadConstraints clause to the deployment.yaml to force replicas to deploy across all the available nodes.

It would appear that the replica distribution behavior is dependent on whether the ingress service is running. If it is not running, replicas are distributed across nodes. If it is running, all the replicas were created on the same node. The topologySpreadConstraints clause overrides this behavior.