Roxie (particularly in the cloud where config often has only 2 or less cpus per agent pod) sometimes logs these msgs per Roxie/pod at a high rate, such as almost 1 million per hour -

ERROR: UdpSender: timed out 1 times (flow=2290419, max=1000, timeout=10, expiryTime=979825468[0] ack(1)) waiting ok_to_send

This is with default timeout increased from 2 ms to 10 ms.

This many msgs of this type strongly suggest not enough cpus or excessive load for given cpu count, but perhaps we can adjust logging logic such that we only log these if the timeout happens > 2 times or > 5 times etc. ?