This was a huge change. I made changes at every stage of the stack and changed the core architecture of my physical infrastructure. I've learned that storage should be the starting point of a new design not an afterthought. The servers are now back at a single location and are clustered into three nodes. The storage provider is Ceph and the storage is clustered amongst the nodes. This will allow me to create new storage classes in Kubernetes to provide resilient storage to deployments. Velero will also be able to take backups from this storage and upload it to Azure. Terraform state is now stored in Azure as well, the Terraform cloud platform was confusing.
This was a huge change. I made changes at every stage of the stack and changed the core architecture of my physical infrastructure. I've learned that storage should be the starting point of a new design not an afterthought. The servers are now back at a single location and are clustered into three nodes. The storage provider is Ceph and the storage is clustered amongst the nodes. This will allow me to create new storage classes in Kubernetes to provide resilient storage to deployments. Velero will also be able to take backups from this storage and upload it to Azure. Terraform state is now stored in Azure as well, the Terraform cloud platform was confusing.