I have fixed this. The blog is now running entirely through PHP's "FastCGI
Process Manager" combined with nginx.
This should make it more stable. If it crashes again, we should file more bugs
(or perhaps just stop hosting the blog ourselves).
As of August 1, the blog sometimes fails to load. This combines poorly with the
fact that we like getting lots of traffic there.
I think that if we stop having Apache's mod_php render the blog, it'll be more
reliable.