A Read Replica Adds Read Capacity Only, and Your Code Must Choose to Use It
This is the node that pays off the single-writer node with the one form of horizontal scale a relational database actually offers, and the curation's outcome cannot be reached without it, since serving real read traffic from one machine is the constraint the whole sub-domain is working around. It is the direct parent of the replication-lag node and the read-after-write node, both of which are consequences of the copy being asynchronous, and it also fixes the boundary against the caching sub-domain: a replica is a copy the database maintains, not one your code maintains.
Scaling Amazon RDS with Read Replicas Using EF Core
Read replicas and multi-availability-zone (multi-AZ) deployment are two complementary strategies for scaling a relational database beyond a single instance's capacity while managing the tradeoff betw…