Postfully Yours

We recently decided to switch our hosting to TextDrive (part of Joyent). Why would we switch our entire system over to a new architecture about a week prior to launch (and, in the process, probably delay our launch at least a few weeks)? Mostly it came down to recognizing what we do best and where we needed help.

The fact is that we aren’t experts in deployment (shocking, I know). Given that we’ve already given up things like sleep, time with friends and family, etc, it wasn’t clear where the time to gain this expertise was going to come from.

We had deployed and prepared our system to launch on a system of fairly heavy-duty linux-based servers. But part of the reason the system needed to be heavy duty is that we had no confidence in being able to scale out quickly. It’s a bad thing when you recognize that you may have troubles if you succeed.

Around this time, I read this blog post (joyeur had been on my feedreader for awhile). 4000 req/s sounded pretty good and having long heard of TextDrive as the premier Rails hosting company, I thought I’d at least put in a request for more information.

After talking to Jason Hoffman, I was ready to go. I had a list of dream features that I wanted to build into our deployment at some point. Jason, in the course of telling me about their accelerator setup (now with more options), basically mentioned everything on my list and told me how it was already built-in to their basic offering. This… pleased me.

I’ll be posting a full review of our experience with Joyent soon. For now, let me just say that it has relieved a huge amount of stress in terms of our future scalability and ability to handle growth. The choice was a couple of weeks of extra work at this point (transition time to the new system) versus having a system unable to grow when we need it. A pretty easy call.