Difficulties of using higher-Q microstrip resonator in oscillator.

I googled few papers, and on provided images there are microstrip oscillators of different configurations (ring, hairpin, split-ring, halfwave) having relative high Q factor. Microstrip width of such oscillators vary about 50 Ohm. Using such microstrip resonators is difficult, i can't obtain 360*n phase and s21 maximum match when simulate manufacturing tolerance. Making microstrip resonator narrower gives lower Q and wider pass-band, then 360*n phase and s21 maximum match obtained easily. Higher-Q microstrip resonator can't be moved after manufacturing, as, for example, DR puck resonator. So how this tuning problem solved in real applications?
The other problem with oscillator designs i found: their microstrip resonators coupled very weak with oscillator (big gaps, etc.). I guess it is because bigger gap gives higher Q-factor. So their S21 for whole oscillator is near failing to start point, because resonator itself provides big consumption of transistor available S21. All that difficulties to obtain a little higher Q?

Resonator Q determines the line width of the oscillator and its noise, too. This is why a high-Q resonators in oscillators are often YIG or dielectric resonators. Their coupling can be optimized easier than with a fixed stripline or coaxial resonator.