Regarding API Gateway database replication, if primary DB fails, secondary DB will process requests from API Gateway.
(1) After that, when the primary DB will recover, will it process requests again by failing back?
(2) As for the failback, what is the time value for the primary DB to fail back?
(1) No. There is not failback function in the API Gateway Master/Master replication. The secondary DB continues to process the existing connection until it will be completed, and the primary DB will process new connection (new requests) after recovering.
(2) There is no such settings as explained above.