Our team uses read-geo redundant storage on Azure and we wanted to be able to snapshot against the primary endpoint and restore from the secondary endpoint of the same account. This is MUCH faster than manually copying the snapshot data from one data center to the other assuming one can do the restore in the data center where the secondary endpoint is located.
I submitted a pull request to elasticsearch-cloud-azure that implements support for restoring from a secondary endpoint and for using more than one storage account.
If you like this idea, comment here and/or add your feedback at https://github.com/elastic/elasticsearch-cloud-azure/pull/93.
Thanks,
Craig.