Partially agree. They should have an accumulated set of tests, so they should by now have known to test connection on different OS, hardware, bluetooth versions. It was mandatory in my professional life to do testing, and whenever there was an unplanned outage, an incident review would look at test plans and update them. So a bug only happens once.
But there is always the bug you’ve never seen.
That is helped to be catched by having a diverse, as deliberately different least-in-common set of beta testers who test on more hardware variety and try things you never thought of.
So good change management, regression testing, customer support, and diverse beta testers, is the leading method.
I get the impression they do insufficient testing and the testers are all very similar to each other.