Asynchronous AG Log Send Queue Growing then flushing quickly

SystemsSQLAdmin 2017-05-03 10:34:26

Morning all,

Have a weird one. On one of our asynchronous servers the log send queue has just started growing and growing. Then the other day it was at 80gb then flushed everything to the log file in a couple of hours. It then had a day of being ok then it started doing it again, then it did the same thing and flushed everything at a random time. Then it started growing again.

SentryOne graph –

My Infrastructure team have had a look and can't see anything that would cause this. Network, CPU and disk all look fine. The only thing that shows up is that CPU and network resource goes really high when the queue finally flushes.

I've had a look at some XEvents and can't see any contention. All I see our some REDO_SIGNAL and REDO_THREAD_PENDING_WORK waits.

Running SQL Server 2012 S3 on both servers.

Anyone had anything like this?