Today’s journal comes from Josh Owens of Transpondr.

After last week’s code rush we decided to start thinking about how to grow a reliable and stable web service. It just so happened that Adam and I were in Portland for RailsConf 08 and we got to have our meetings in person on the best approach for something like this. After about an hour of discussion we came up with a multi-server, multi-app setup that should allow us to survive server downtime. Further, phased code deploys should allow us to ensure we have smooth roll-outs for new code.

Transpondr

A little serendipity also occurred this week, while in Portland. We ran across an awesome new service called New Relic (http://newrelic.com) that digs deep into Rails internals and monitors what your app is doing. This great tool gives us a ton of insight into how long it takes various pieces of your rails app to run. I took an hour going over the app and tuning everything to run much faster. I would highly recommend finding a tool like this (if you don’t use Rails) or using New Relic.

How To Split An Atom’s Startup Reactor has been brought to you in part by the following sponsors -

ThemBid
IERSolutions

(Image) (RSS)