So … here’s what happened, according to David Karp.

staff:

Our engineering processes seriously failed this afternoon and cost you and your blogs nearly 4 hours of downtime and almost 100 million views.

Painfully, this isn’t the first time this winter I’ve had to give you similar news.

When incidents like this happen, our entire engineering team comes online to support the recovery as needed. Immediately after, we begin taking every measure to protect from the uncovered issue in the future.

We are constantly working to shore up our processes and solidify the stability of this quickly growing network, even more so as we’ve fallen behind the last few weeks.

Tumblr’s success is supporting your success, and we take this mission very seriously.

This was a rough one. As Hacker News points out, they had to shut off the DNS to fix it.