My experience:
- make a ticket with email admin support
- get it closed
- make another ticket
- get it closed
- find other channels to backlevel email support and supply exact datetime
- then they will fix loadbalaning/delegation/… whatever
In my case they found 1 bad server out of “many” by looking at logs and logs and logs.
In the meanwhile I did dependencies up- and downgrades, retry-scopes en masse and anything to keep my process throughput somewhat stable.
The rootcause was fixed in mail infrastructure. Until then I was shooting at angry birds, figuratively speaking…