Just a quick follow-up post on how things went after the last post about this topic
So after finding that scaling the webservers up wasn’t the solution to the problems we experienced with the webservers, we started to look elsewhere, and actually found that a load balancer that was set to do round-robin load balancing instead of least used, caused all the hiccups. Once that was changed we started hitting the limits on bandwidth instead.
And may 14th the website was actually launched, and there hasn’t been any stress at all on the infrastructure since launch.
Here is a shot of how everything looks after launch.
The spikes around may 1st. is due to us stress testing the system before launch. And there is a tiny little spike on may 14th when the website went live. And there hasn’t even been anything that vCops found to be outside normal (no yellow boxes) after the website was launched.
Overall I’m really pleased with the fact that vCenter Operations Manager stores your performance stats for you to review at your pleasure, and the amounts of metrics you can pull out of the system is really amazing.