Background: Our domain page.et is not accessible by Google's mobile-friendly checking tool and search console. The same seems to be true for all other .et domains I tested.
The reason is not the robots.txt. Google bot does not even try to make a connection to the server (verified via tcpdump)
I want to understand the technical background of the following answer to another user with a similar / same problem:
(From Russian via Google Translate)
Hello Alexander ! At the moment, I only learned that the problem is on the side of the .et nameservers - they "block" Googlebot Accordingly, neither you nor Google, at the moment, have the opportunity to directly influence the solution of this issue. Google specialists are trying to contact representatives of NS .et to resolve the current situation Unfortunately, there are no guarantees / messages about the time when this issue will be resolved.
So how can a nameserver / zone "block" a Google bot? And how can such a block be verified?
The domain can be resolved by Google's own nameserver 8.8.8.8 (verified by dig page.et @8.8.8.8) and I would believe, that the Google bot uses the same nameserver
Remark: I also posted in the Webmasters Google Support forum but here I would like to understand the technical background how a name server can block Google bot or any other HTTP client from accessing a site.
Putting it simply, Google or any other name servers are going to have to recurse to .et nameservers in order to access any information in them; if they collectively block Google, then any URL-based requests will fail, as name resolution will fail.
This is an overly simplistic answer of how, but it’s possible.