Questions on Kong Internal time variables

Cross posting from Gitter in case someone ever wants to search this:

Anyone know the actual difference between KONG_WAITING_TIME and KONG_RECIEVE_TIME? My notes say waiting time is the time it took for the backend to process the request, and my notes on receive time are the time it took before Kong fully received all headers and response body from backend. My notes must be lacking detail cause these sound kinda identical in my mind haha.
Some logs show waiting time 790 ms and receive time 119 ms, these seem so far apart that my notes definitions must be off.

Or does that mean maybe the backend finished its bg processing in 890ms and took 119ms to actually deliver the payload fully back 2 kong so kinda a total time of 890 +119 gives you the full time before a client would start getting it back from Kong/Nginx(well maybe not true cause I suppose during the 119ms window kong is already sending headers and response data back towards client right?

I think that makes more logical sense, because on some tx’es of 1.2mb responses I will see waiting 7 seconds then a recieve time of 7 seconds too leading to like 14 seconds of latency which is not good times, but those 7 seconds of recieve are not usally kongs fault I would think. Although its interesting the HTTP log plugin bunches the KONG_RECIEVE_TIME element in with the other Kong latency vars, making the kong = (latency) value super high when this extra 7 seconds of slowness may not really be kongs fault per say but more just the time for backend after processing to push bytes across the wire to kong, maybe representing poor network bandwidth behavior? Might give the false impression Kong was “slow” on the transaction eh?


© 2019 Kong Inc.    Terms  •  Privacy  •  FAQ