As much as we all think google is smart at processing everything in a perfect manner, I believe there are still some things it can't do, and I think one of them is establishing a secure connection just to download a plain text file.
What I would suggest trying in order to make every single search engine happy is to make your robots.txt accessible without requiring a secure connection to be established first.
When I tried accessing:
http://soundhaus-band.de/robots.txt
I get a redirect to:
https://soundhaus-band.de/robots.txt
This can throw search engines off because they're looking for robots.txt content only to find a redirect.
You should make http://soundhaus-band.de/robots.txt produce the exact same output as https://soundhaus-band.de/robots.txt to make everyone happy.
If that doesn't work, then send google negative feedback about their robots.txt detection service.
http://...
property instead of thehttps://...
one? "fetch the robots.txt as Google" - yeah, don't do that, it's not a file you want indexed.robots.txt
a few days ago (on 01/11/2016). However, it was just a spike on that day (which coincided with a server update).