Setting up production and disaster recovery clusters with replication

Hi Team,
I am reaching out to gather information and guidance on setting up two clusters for our organization’s data management needs. We plan to establish these clusters in two different data centers, serving distinct roles: one as our primary production cluster and the other as a disaster recovery (DR) cluster. Our objective is to ensure high availability and robust data protection.

What’s the expected replication volume in production?

Hi Alexey,

it will depend on the replication sync interval
so we can say the expected event per second will be from 1000 to 3000 and each event can have an average size from 1 to 2 kb
so the expected volume in average will be 6000 kb/second

is that constan load like 24/7 or peak time ?

it will look like 24/7

You could try with

the other way , depending on your recovery scenario and time to recovery would be to set up a Read Only Replica in the other region .

Replicator can only get to 1K+ events per sec when used with partitions. So, it won’t be able to keep the global events order, only per-stream order. It can, in theory, handle higher load if the number of streams is low. Like, when you just have one stream, it will go fast.

Is there any guide I can follow for configuring Guinness for using Replicator with two on-premises clusters?

@m.saied : Configuration | Replicator

you can set up the source & sink via the onnection strings