NimbostratusMy website is slowing down on peek load. I would need to find
1) How much time ltm (virtual server) is taking to process the traffic before it is passing to http pool 2) Pool response time 3) Is there a way to compare historical performance peak load vs non-peak load ?
Any help would be appreciated.
Cirrostratusyou can compare those with tools like tcpdump tarce or fiddler tools...
NimbostratusHere, i can suggest you few steps, which may help you to nail it down to an exact issue. 1. such performance issue with any other application hosted on same F5 reported?
if you know the ip addresses of real servers used in VIP, try to connect them in your Intranet environment, and see if you observe something.
sometimes we may need to use the load balancing algorithms, may be least connections (member) may work better as per the application instead of round robin algorithm.
if these servers are like shared, i mean these servers are serving requests and connections to other applications, then i would suggest using one connect profile. which will decrease the load on real servers.
May be if real servers area loaded reasonability, then approach adding few more physical or virtual machines into the VIP.
you can take advantage of SCCM tools which can also give you good performance statistics of system resources on these real servers.
--Ravi K. Malhotra--