Hi @Ahmed_Dhaoui ,
I have not worked with it yet directly, but I can say that, in theory, it should work for as long as you keep attention to some details.
For as long as you implement High Availability, it should work fine: https://aws.amazon.com/blogs/apn/high-availability-and-observability-using-red-hat-advanced-cluster-...
The idea is similar to Kubernetes Kosmos released by Scaleway or other solutions: https://www.scaleway.com/en/kubernetes-kosmos/
In short, you will have a split design of 1 K8s cluster for control, administration and load balancing of your different clusters, and then you have your separate clusters and namespaces. So in this sense, that technology will take care of all for you.
I mentioned there are some details to keep in mind, and it all falls down to the backups, state of storage/memory and what is synchronized between the different kinds and locations of storage on the clusters. The images/containers/namespaces should be replicated just fine. But the Persistent Volumes (local & Shared Storages) will need to ensure:
1- all memory data to be frequently backup (SAS Backup)
2- run periodically data integrity checks to ensure no broken or locked files and all files been properly synchronized across all storages
Little disclaimer here: Please do keep in mind that for as long as it has not been tested by SAS Institute and not documented, it will be not officially supported as such.
