Consider the following scenario: I have got a couchbase database installed in 3 different clusters. As you may know, couchbase db offers seamless selfreplication between clusters.
I would like to know, if I push some records into a cluster db, how long does it take to be replicated to the other clusters?
An original approach I am taking is:
- Have a writer client injecting records into a cluster only.
- have a set of clients retrieving data by connecting to the other cluster(s), they have to be constantly querying the database every second. Is this approach ok, or would you think of a better use case? I know the delay between client reads would be an overhead that will affect an exact calculation.