A determination is made that a relational database management system (RDBMS) is configured as a distributed availability group. The distributed availability group spans first and second availability groups. Each availability group includes a cluster of servers hosting replicas of a database. One of the first or second availability groups functions as a primary availability group. Another of the first or second availability groups functions as a secondary availability group that is available as a failover target should the primary availability group become unavailable. A name of the distributed availability group is obtained. A first server in the first availability group is directed to backup a replica of the database being hosted by the first server. The directing includes instructing the first server to index the backup against the name of the distributed availability group.