ReplicationLatency might vary for different source and destination pairs, depending on how geographically far apart the tables are from each other. For example, ReplicationLatency is typically lower for global tables in the Europe (Ireland) and Europe (London) Regions than for tables in the Europe (Ireland) and Asia Pacific (Sydney) Regions.
ReplicationLatency and PendingReplicationCount might increase when an AWS Region becomes degraded. Use the AWS Service Health Dashboard to check for events in the Regions where you have replica tables. If there's an event, then you can temporarily redirect your application's read and write activity to a different Region.