Geographically distributed datacenter replication?

I’m not sure about this feature in the 1.0 roadmap. Release 21.03 adds something called learner nodes, which are nodes that are able to connect to a Dgraph cluster and replicate data, but do not vote on proposal.

When combined with best effort queries (and disabling mutations on that node), you could use such nodes as a read replica.

I don’t have a timeline for this on Cloud.