I would like to use dgraph as a distributed key-value store to replace things like etcd or cassandra or consul kv. I appears to have comparable or better high availability and horizontal scalability. Is anyone doing this in a production environment?
funny you ask, I have been thinking how great a dgraph backed k8s would be (replacing etcd). Mostly because of dgraph’s operational features (distributed RAFT, backup/restore, etc.) and my expertise running dgraph vs etcd.
Currently I run k3s->kine->haproxy->cockroachdbx3 and I am not proud of it but I get great operational features.