so not sure if this issue has been addressed or not, but...
I noticed that we have a daily latency with our application calling on published services, this only occurs with the first call on services each day. once that first call on services has been made it seems the system is finally "awake" and running properly.
has anyone else experience any similar issues? if so is there a proper way to resolve this daily latency?
we are calling on both a map service and a handful of geocoding services...
they all show:
min # of instances per machine 1
max # of instances per machine 2
I'm thinking it has something to do with the Recycle settings that run every 24hrs, its seems once that has run that the first call on services lags, then the services are fully awake after that first call. its about a 15-20sec lag on the first call on services, once that call is complete any calls after that are almost instantaneous.
My guess is we probably have the same lag after the daily recycle, but I haven't noticed or been notified.
Any thought of trying to schedule a "wake up" app after the recycle so they are primed for the day. Right now I can't think of a way to do that, and maybe you already thought of that, but if not...
It will most likely come to that!
I guess my last question for anyone here first would be if this is something to be expected if the client is going through a Load Balancer to access the services?
a) I wonder if that in itself would cause a delay from time to time.
i just wanted to provide an update on this situation...
1) we instituted some automated testing in our environment
a) 1000s, if not 10000s calls to services
We found something VERY interesting happening, and maybe others have seen this:
both tests were run within the same time frame, same environment, to the same server instance:
Test 1 - VM - showed a 3sec latency on a return of custom fields from a service (95% of calls experienced latency)
Test 2 - Physical Machine - showed no latency at all on return of custom fields (0% of calls experienced latency)
has anyone else seen this? where a VM somehow causes a latency factor or at least seems to be the culprit.
thats where we are for now.