I have a question about the Rancher cluster itself. Would it be possible to spread the cluster over multiple environments? We have our on-prem production datacentre, a DR centre and presence in Azure via an Express Route whith <20ms latency to production (can't remember if 5ms or 12ms, will check with networks team). Would it be possible to spread Rancher's local cluster across these points, to allow it to remain active, if any one of them should experience an issue? Or am I whistling Dixie through parts unmentioned? Edit: layout

Yes, tho etcd is highly disk-latency dependent. I run Rancher Server HA across two datacenters in the same geographic region with no issues unless I lose some internet connectivity.

Thanks for the info. This is very helpful.