A GoldenGate configuration consists of multiple processes each of which perform specific functions. Even though these processes are loosely coupled and maintain their respective checkpoints, an issue with one of them can cause the overall replication for that particular stream to fail. This behavior is not confined to just an issue that will cause a failure, it also persists to an issue which causes delay to one of the processes. So, if the Extract process is overloaded and there is a delay in reading data from the redo stream, the downstream Datapump and Replicat processes will receive the data only at that speed. Although monitoring the lag of an individual process can certainly highlight the sudden spikes in lag times, it will not provide a breakdown of what stage of the replication is causing the majority of the lag. This is where you need to implement a technique called Heartbeat . In this recipe, we will look...