On the morning of October 1st, we suffered a 10 minute downtime due to a runaway process on our primary webhosting machine. This slowness caused roughly 100% of requests generated during this time to be completed out of SLA.
one of the services running on the machine crashed
fixed the cause for the service that crashed and then restarted it
css2inliner consumed too much memory
all customers were affected by this