Would love to hear about what it took to transition to AWS. Any chance Joel would divulge some of the driving factors (reliability being one)? How about cost? Unlike my paltry webapps, trello gets a lot of activity and no current revenue stream, so I'd love to know more on what the cost difference is between running your own vs in the cloud. How long did it take to move over? Are the old servers being used as backups in the event Amazon has their very rare aws outage?
I know that's a lot of questions, but heck it would be an article I'd love to read and something I believe would be very relevant to the community as well!
> Are the old servers being used as backups in the event Amazon has their very rare aws outage?
AWS is essentially acting as a backup of the old servers, as the datacenter is in danger of losing power.
That said, the cost, reliability, etc. had been carefully thought out, albeit as a somewhat longer-term project than it ended up being. More info will certainly wait at least until the current state of things is cleared up - Fogbugz and Kiln are still in the datacenter.
Thanks for the info. I know things must be pretty crazy right now. Quite literally in the trenches. Just goes to show what an amazing team behind an amazing product can do. Kudos!
I know that's a lot of questions, but heck it would be an article I'd love to read and something I believe would be very relevant to the community as well!