Over the weekend we made a bunch of bug fixes and improvements to Twitter which resulted in more reliability in message delivery—this increased overall traffic.
Ordinarily, increased reliability and traffic is a good thing. However, this comprehensive delivery of messages together with two severe cases of abuse strained Twitter’s primary database enough to degrade performance.
Today we worked to address these performance and reliability issues. We also took notes and preventative measures so Twitter will be smarter about this stuff next time.
The web site is feeling snappy and back to normal which means we will now be investigating and finding a fix for the SMS issues which some folks have been reporting.
It doesn’t always feel like it while we’re digging in and making small improvements but we are building a better, more reliable service every time we fix a bug.