Fix logging of replication events processing.

We should have an explicit markers of replication event processing start and stop with the final state of the event. It will help understand if processing of replication event was done properly or some interruption happened as well as identify events that take too long to be processed. The starting log entry includes event as after successful processing it will be removed and there is no way to check its target/source nodes and repository. All event entries are bound by 'correlation_id' so it is easy to get all related log entries by scanning for it.

Logger is propagated throw context that is passed throw the call chain. It allows to add additional logging data into the logging entry without passing logger explicitly to each method/function.

Part of: #3047 (closed)

The current situation with logging of start and stop of the replication: It is almost impossible to identify correlation between start and finish log entries of the same event (now we can as replication process is sequential for each praefect instance, but it will be changed once we begin process multiple replication event simultaneously).