You guys are smart. Three things: 1) The cache is slow to warm up. Therefore there was some latency involved with the restart of the service.
2) Some of you may have noticed problems about an hour after the restart. Some important objects expire from cache after an hour. Since there was a huge influx of objects expiring 1 hour after the restart presumably together there were problems as the database began to be overworked. We're doing a post-mortem now to determine how to better keep the cache hot even after objects are cached nearly simultaneously after restarts like we experienced today. 3) The downtime today was to increase capacity, which does not exactly translate to performance. Thanks, Doug -- Doug Williams Twitter Platform Support http://twitter.com/dougw On Fri, May 8, 2009 at 5:28 PM, Nick Arnett <nick.arn...@gmail.com> wrote: > > > On Fri, May 8, 2009 at 4:11 PM, Eric Blair <eric.s.bl...@gmail.com> wrote: > >> >> Another maintenance-related question - are things completely back up and >> running? I'm asking because I'm seeing inconsistent behavior. >> >> In my servers logs, I'm seeing a number of timeouts when talking to >> Twitter. Also, when I try to run curl commands from my servers to Twitter, >> they are _extremely_ slow, sometimes timing out. Some of my servers are >> systems that regularly communicate with Twitter while others rarely do so. > > > Same here. > > Nick >