Yesterday around 12:30pm GMT (see in your timezone) we experienced a network outage. As information came to hand, we updated details on our status blog. Over the course of two hours, our network connection to the Internet went up and down, causing FastMail to be intermittently accessible and inaccessible for those two hours.
We've worked with NYI to identify the problem, which turned out to be a faulty supervisor card in the upstream switch we were connected to. The intermittent access was caused by switching network switches while troubleshooting the issue.
The faulty switch has now been replaced, but to avoid the possibility of this problem happening in the future, we're working with NYI to create a dual/redundant uplink for our configuration.
Over the past couple of years, we've been striving hard to create an extremely reliable replicated infrastructure. This setup has been very successful, and creating dual/redundant uplinks will be another step along this process.