Read Replication but not round robin.

[jatinkhurana1998]

So the read replica we use have a time out of 2 seconds, and for one of the search queries where we send like = %xyz% the indexes of the table doesn't work and the query gives a timeout.

I need to direct these request to a different replica which have a larger time-out time.

As even if we have what current support we have of putting more replicas in the the file, requests will be served in a round robin fashion, and we won't be able to send these specific requests to that replica.

If a feature like this exist, please help.
Thankyou.