Send PM
Followers 349
Posts 9916
Boards Moderated 9
Alias Born 05/25/2003

Re: Phil(Hot Rod Chevy) post# 128523

Saturday, 06/13/2009 12:05:52 PM

Saturday, June 13, 2009 12:05:52 PM

Post# of 205762
WEB3 and WEB4 are effectively identical servers.

That's not where the problem is - the cause of the latency has shifted from the traffic overwhelming one webserver to overwhelming a load balancer that is spreading the traffic over two webservers.

The "good" underlying cause is combination of both a steady increase in overall traffic and a corresponding increase of numerous bots and crawlers that show up during peak traffic periods and read hundreds of posts per minute, with sometimes more than one running at the same time.

The good news is that we know where the weak link is. It's going to take a little time to upgrade to a load balancer that can handle this and even higher levels of traffic. Yesterday we peaked at shipping nearly 55mb per second, most of which comes from dynamically generated pages created at the time of the request. That's quite a demand, especially considering that just a few weeks ago we were peaking around 15mb.

We'll get it fixed and appreciate everyone's patience.