Wow. We had a feeling people would love the EtherPad product, but we had no idea that our traffic would grow so rapidly just from the initial launch. Thanks to all the enthusiastic users who submitted great feedback and helped to report connection problems! We read each and every one of the thousands of messages, and we sincerely appreciate them.

In short, we got way more traffic than we anticipated [1]. We will be able to keep up with the growth in the long term, but the initial surge of interest has been wreaking havoc on our servers. We ran load tests before we launched, but we didn't simulate the systematic effects of thousands of simultaneous typers sustained over days.

It's been exciting keeping our servers alive for the past few days. We've been sleeping in shifts, and Aaron wrote a script to play the Dr. Octagon song "Technical Difficulties" through a boom box at full volume every time the site went down. But working in that interrupt-driven mode left no time to resolve some of the bigger unexpected architectural issues required to provide a good experience for many thousands of simultaneous typers.

To address these issues, we're going to take a step back for a moment and restrict new pad creation to a beta sign-up list. (Existing pads will continue to work for anyone). This will let us focus on fixing core problems instead of fighting fires.

We appreciate your patience and hope to have the site back open to the public in short order. In the meantime, you can sign up for the beta or follow our mini progress updates on the appjet twitter feed.

With Warm Regards,

The AppJet Team

--

[1]

Perhaps if we thought more about it, we could have guessed this would happen. A site like EtherPad has what marketers call a "viral" growth pattern: it's more useful to you if you invite your friends, and then it's more useful to them when the do the same, etc. 80% of all people who visited a pad got there from a direct link or email invitation, suggesting that people were indeed sharing pads with others. On the other hand, you never know before you launch a product quite how much interest there will be.

blog comments powered by Disqus