This weekend we updated our DNS to a different ISP. We made the change Friday night around 6 pm and it updated Saturday night around 6 pm. Emails are coming through when it is slow, but during the peak hours it slows down. Even internal emails are taking 1-2 hours to arrive. I do not know what to look for to see what is causing the problems. There are about 4000 emails in the Queue Viewer on the Exchange Management Console.
Just for your information... we use a fatpipe system to bond three different ISP lines. Traffic is aggregated depending on which line has the best latency.
When we were making the change we removed the DNS forwarders on the exchange server that went to our old line, and added the DNS info for the new line. It seems to me that we would need all 3 DNS information on the exchange server, does that seem right?
Does anyone know what I should try to determine what is causing the queuing? Any help is greatly appreciated.
Viewing an individual message in the Queue Viewer should show you which error code was returned when delivery was attempted.
For DNS, since your traffic may come out ANY pipe, and oftentimes ISPs will only allow queries to their recursive resolvers that originate from their OWN pipe, this could be a likely culprit for why message delivery is intermittent (successful when the new ISP's pipe is used, failure when one of the others is used).
I'd suggest changing over to an open, global DNS resolver; Google runs some that are solid, for instance.