Metrics for RTOs in TCP FLOW


#1

In the API doc for Flow.server.rto and Flow.client.rto, it states this metric is the “number of RTO out when the device is acting as the server” or “…acting as the client”.
I displayed with DEBUG this metric in a trigger attached to a device that I am seeing a high number of RTOs on in the “L4 TCP DETAILS” screen for the device.
I see numbers in the DEBUG log ranging from 0 to 30 or more. (Mostly 1 or 2)
What do these numbers mean?
As I understand it from reading the info on the forum, an RTO is a stall of 1 second in the flow as the sender stops sending because of too many retransmissions. So, are the numbers I’m seeing the times that the flow is stalling for a second.
Thanks.


#2

Correct. The RTO number represents the count of number of times the RTO event happened on the flow between the client and server. RTO is typically caused by network congestion and results in a stall of at least 1 second but could also be more. Hope this helps.
Thanks