Service Fabric VMSS Nodes Running Out of Memory - Scaling Up Considerations

Hello,

We have a Service Fabric cluster configured with a bronze durability tier, consisting of 2 frontend nodes and 7 backend nodes. The VMSSs are utilizing the Mirantis image "windows_with_mirantis_container_runtime_2019."

Currently, we're facing an issue that the backend nodes running out of memory.

Several questions need to be addressed for determining the best approach to resolve the memory issue:

1 - Do we need to to create new VMSSs and redeploy the Service Fabric cluster to VMSSs with more memory, or is there a simpler interim solution to mitigate the memory issue before considering migration?

2 - The environment is deployed using ARM templates (dev, test, prod), but code drifts have occurred since the last deployment. What is the best way to identify the drift? Currently, the "- what if" parameter is being used, but the output is challenging to interpret and shows a lot of false positives.

3 - Is there any other OS image with container runtime available for use? The current Mirantis custom image is not supported by the update manager and we have to perform OS updates manually at the moment.