Den Support habe ich aufgegeben. Man muss jedes Mal Handstände machen um von extern die Hoster-Infra bei Problemen genaustens zu analysieren, mal abgesehen von Späßen wie dem Rettungssystem was richtig Laune macht bei Produktiven Maschinen.
Selbst wenn man das macht gibt es nur dumme Textbausteine. Das Problem auszusitzen erscheint mir seit einem Jahr sinnvoller als meine Zeit zu verschwenden
You summed up my experience. 
When you're asked to boot into the recovery system to test something on a production server.
And what about when there's a network issue and instability occurs at random, rapid moments, and you're asked to test MTR on the recovery system? Hahaha.
Imagine leaving the server offline in the recovery system until the problem registers it again? Total madness.
I was having a packet loss issue between Netcup servers in the same datacenter.
I logged the issue in Ubuntu's own MTR, but they wanted me to boot into recovery mode to perform the same test.
Note that the problem occurred at random times and was always rapid.
To give time for registration in recovery mode, I had to leave the server in recovery mode, running MTR until the problem was registered. This was on a production server. Sometimes it took more than 10 hours for the problem to occur. HAHAHA.
Luckily, I was able to log the issue in the recovery system. The problem was that it had already been detected and logged in Ubuntu itself.
