The amazon AWS services have been extremely popular with lots of startups and has received very positive reviews. AWS is a way to leverage the Amazon infrastructure to provide lots of scalability for image hosting, CPU tasks, DB and more.

So what happens when an outage hits Amazon as we saw today ?

Impact wise, what struck me this morning, were the vast number of broken images across the web ( from sites that rely heavily on Amazon S3 ) and twitters of people agonizing about their businesses effectively being shut down.

A post on the Amazon forums is a good example:

I have to add a major ME TOO here. My business is effectively closed right now because Amazon did something wrong. I’ll have to reconsider using the service now.

For us on WordPress.com, we use S3, but only for a small portion of the image serving, and thanks to our architecture, were able to automatically deal with the outage this morning with no impact to our users. Our systems wrangler Barry has a post about it:

Currently we serve about 1500 image requests per second across WordPress.com. About 80-100 per second are served through S3; the rest being served from our local caches. When the outage occurred, our systems detected the errors and automatically sent the requests normally bound for S3 to local image servers that we use for backup and failover purposes.

Read Barry’s full post