[pjscott]

The site is running on Wordpress. Incidentally, one of the major reasons for Wordpress blogs to go down under heavy traffic is excessive keep-alive times causing too many Apache threads to be open. It's exactly this lots-of-connections problem that node specializes in handling gracefully.

[pjscott]

A postscript to this, now that the blog is back up and I can do some measurements: the server is Apache 2.2.14, and it's using the default keep-alive timeout value of 15 seconds. Looks like a typical case of Wordpress and Apache with the default configuration suddenly getting slashdotted. More on this from patio11:

http://www.kalzumeus.com/2010/06/19/running-apache-on-a-memo...

[stellal]

That was embarrassing. I do know how to craft a good apache config, and am just going to claim a case of being stupid.

It is ironic that node is in fact designed to handle this type of memory/connection scalability problem reliably, but in this case it was pure administrative (me) error.

[_pdeschen]

Throw nginx in front :-)

http://blog.rassemblr.com/2011/01/wordpress-need-for-speed-o...

[allan_]

omg excessive keep-alive times