News for hackers. Stuff that matters.
Re:Github service outage by ChuckMcM on Friday November 30, 2012 @04:41PM (#4855612)
Ouch, this bit "Features are sort of irrelevant compared to recurring downtime." is key. If you're going to build a network based business, you have to invest in being operationally nimble. Just because DevOps isn't 'sexy' doesn't mean it isn't important :-) (and no I'm not saying that just because I am an operations guy, but I feel the pain when looking for DevOps engineers)

I also feel GitHub's pain. We're constantly dealing with things that come in over the transom at Blekko and getting a constant stream of sales guys and gals trying to tell us how their product will solve all our issues, and as I tell folks, if you don't know how to solve the problem how is it that you trust this solution being offered is sufficient? I've known people who will respond, "Well if it isn't we'll just beat on them until it is!" but that doesn't help the 'gurgeous's of the world who are trying to use your service while your vendor it getting beaten about the head and shoulders with a tuna.

I've advised a couple of startup founders who had under 'scaling' in their proposals "EC2/S3 totally scales" that while it was true, lacquer oil paints create masterpiece paintings, except if you don't know why a masterpiece is a masterpiece, just having the paint won't get you there.

Re:Github service outage by gurgeous on Friday November 30, 2012 @05:07PM (#4855720)
I agree completely. I've had to deal with scaling challenges before, most memorably at Urbanspoon. I'm not judging anyone. This stuff is just hard.

As an entrepreneur/engineer my day is already jam packed. We have a tiny, full-stack engineering team here. I really just don't have time to babysit services that suffer from significant downtime. Certainly I could mitigate the issue by distributing our repos to multiple places, self-hosting, etc. but I'd prefer to just use something reliable.