So, that wasn’t fun, yesterday:

We were also hit, given we run DNSSEC on our resolvers. Interesting some large open resolver operators use Negative TA’s for this sort of thing. Not sure how this helps with the DNSSEC objective, but given the kind of pain mistakes like these can cause, I can see why they may lean on NTA’s. Mark.

We did not use an NTA, but we did flush our cache immediately once Slack had fixed their problem. I think that’s the right balance of carrot and stick.


Tend to agree with this approach.

But I can see how an issue like this could be potentially religious. DNSSEC deployment rate is bad enough, as it is.


Friday is always a good day to do such change. :smiley: