Additional logging when dealing with delayed slaves

When detect_replication_lag setting is used, it would be great to have a way to get the following information recorded in Maxscale log:

1) Seconds behind master (or exact delay) as detected by Maxscale when there is a lag
2) Time when read connections are routed to Master because of delay
3) Time when read connections are routed back to slave (they are within allowed delay)