[elliottcable]

Let’s compare:

http://twitter.com/tumblr/status/11570891053862912 … 130 characters, 21 words, no actual information other than that which could be assumed: we’re working on it (can be assumed), we’re sorry (should be hoped).

https://github.com/blog/744-today-s-outage … ~2,000 characters, 335 words, and an absolute plethora of transparency and information. In fact, more than we probably needed to know, even the techies among us… but nonetheless, quite reassuring and friendly on GitHub’s part.

I’m sorry, Tumblr. You’re failing. Not technically; nobody cares about a little downtime, you’re not Google, you don’t make aircraft stay in the air, you don’t track weather patterns that could tear apart an entire coast, the world’s not going to end. No. What you’re failing at is being that which it is absolutely most imperative for you to be, being which itself being the single impetus for your existence: a friendly, transparent, communicative tech entity that your users can trust.

[harryh]

I don't think it's fair to compare those two. One is a while it's still happening tweet, the other is a post mortem from after the incident.

Tumblr's engineers are likely very busy right now working on getting things back up and running. You can't expect them to take time out from that to write a detailed incident report for the public. Wait and see what they release tomorrow before making a judgement.

[elliottcable]

Simply put, what mcdavis said. GitHub also has a history of being communicative during issues; not to mention near-instantaneous fixes of problems (anybody remember how unbelievably quickly they re-architected their entire site to deal with the scaling and speed problems that had barely started to surface? how about the lack of speed issues since?) as opposed to these things I’m hearing of people seeing the same, obvious, bugs over the last two years or so.

[mcdavis]

If Tumblr typically (or ever) posted posted post mortems, you'd be right. Usually, it's odd for them to tweet about issues.