WarpStream Multi-Region Clusters guarantee zero data loss (RPO=0) out of the box with zero additional operational overhead. Multi-region consensus and automatic failover handling.
The data and control plane both are made RPO=0 independently. For the former, you configure the Agents to write to multiple object storage buckets across more than one region (this ensures you have at least one copy of data if a region goes down). For the latter, that’s managed by WarpStream, so you just select a multi-region control plane.
You would deploy a single cluster across all regions. It wouldn’t really be an active-active architecture (it would be a primary-secondary architecture), as each topic-partition will have a leader responsible for writes, and that leader will only live in a single region. If the leader goes down, a new region and leader will need to be selected. The networking setup is complex, and you have to be mindful of leader election and KRaft settings. It will be really expensive due to networking fees.
The feature is available on all Enterprise tier WarpStream clusters (see our pricing page for differences between WarpStream tiers). Contact us to discuss pricing.