This one is especially popular with databases, and it helps with scaling them, too. Replication means you will run a few instances of your service in parallel with duplicated data, all handling incoming traffic.
Don't confuse replication with sharding. The latter doesn't require any data redundancy, but can often bring you great performance at scale. If you're using Postgres, we recommend you try out Citus (https://www.citusdata.com).
In terms of databases, there are two ways you can replicate.
