In Linux, you will need to go through a lot of documentation and forum until you learn enough about the component to understand how to fix the issue.
This may seems a huge effort, but it is very rewarding: your problem is solved forever, you know what caused it, you have discovered many useful features of you OS that you did not know, you have learned how to analyse similar issues.
Linux is very stable: almost everything I have learned 20 years ago remain useful.
Skills with Linux tools help diagnose the issue, but they don't always point to an obvious solution. As I mentioned earlier, I often have issues with sound in Linux, especially when trying something beyond the stock configuration (e.g. JACK interfacing with PulseAudio and ALSA). The graphics stack in Linux is slowly improving, I don't see why the same can't be done for audio. RTFM shouldn't be a requirement for basic functionality (graphics, sound, input, networking).
I've used various Linux distos over the last 10+ years. If you use something like Slackware, it is rock solid. I still have some 5+ year old installs that simply just work. However, distributions that update/upgrade and install software automatically have lower stability as more people are involved in creating packages, and not of all of them are skilled or careful enough. Packages are often patched versions of vanilla software and sometimes the patch was just wrong because it does not sync well with some other installed software. I still recall major mess by Ubuntu and Fedora and their derivatives at many points in the past.
I've used various versions of Linux and Windows, and in my experience Linux is great to use, until something breaks and you want to refresh that component. When I first started using it around 2005 it was network issues (luckily things are much improved now), when I use it now I most frequently have issues with sound. On the other hand, Windows might not be as nice a development environment, but at least you can rely on the drivers (for the most part).