[Question posted by a user on YugabyteDB Community Slack]
Looking at the docs, https://docs.yugabyte.com/latest/explore/multi-region-deployments/ - over here it says that the xCluster pattern leads to some data loss (RPO). What are the specific scenarios when data loss can occur?
The scenarios where XCluster would have data loss are in write-write configurations. Two options are [1] conflict resolution where last writer wins and [2] Read-Modify-Write when reading data that has been modified by the other cluster but is part of outstanding replication transactions in the asynchronous lag.