1. Check your group’s instance strategy.
   • If your virtual node group has more on‑demand instances than defined, extra instances return to spot instances when spot capacity is available. This behavior is called the fix strategy.
     
     
2. Review the log message. You may see something like:
   • Replacement of type Out of strategy has been canceled. Reason: Instance contains stand‑alone tasks, and the group’s configuration doesn’t allow termination of stand‑alone tasks.
     
     
3. Update your group configuration so the autoscaler can terminate standalone tasks.
   • In the API or in the cluster’s JSON file, set:
     "shouldScaleDownNonServiceTasks": true 
     
     
4.  Save the changes and allow the autoscaler to retry the replacement.
   
   

The standalone task and the instance are terminated. The standalone task isn’t redeployed because it wasn’t created as part of a service.