What Is the Time Delay for Primary/Standby Replication?

When standby instances cannot keep up with the updates on the primary, this generates replication delay. If the standby SQL and I/O thread are running, the replication delay is a positive value measured in seconds. If the standby SQL thread is not running, or if the SQL thread has consumed all of the relay log and the standby I/O thread is running, it is NULL (undefined or unknown).

The delay for primary/standby replication cannot be calculated using a formula as the delay is affected by the following factors:

• Network communication status
• Transaction workload on the primary DB instance in transactions per second (TPS)
• The size of the transaction executed by the primary DB instance (this affects the duration of transaction executions)
• Load balancing of the standby DB instance and read replicas
• If the recovery model is set to Simple, the primary/standby relationship cannot be established and the maximum value of the Replication Delay metric will be reported.

If the primary DB instance has a heavy load for a certain period of time and executes a large number of transactions per second, replication to the standby DB instance will be delayed. This delay is generally a few seconds.

RDS for SQL Server: To check data consistency between the primary and standby DB instances, view Replication Delay on the Cloud Eye console to obtain the value of the primary/standby replication delay.