13.56 x 2 Msamples/s is the absolute minimum you need to reconstruct the signal with a sin(x)/x approximation.
In reality you need at least four times, up to 10 times the shannon limit. Why? because mostly you seek for non-linearities or higher band harmonics.
And yes 32k (in fact 14k for DSO from hantek) is small, but it’s enought to cover 1400 15Mhz cycles at 10x sampling. Learn how to generate triggers, even with high end agilent scope I have to generate accurate triggers to capture small event and navigate trough long events.
And hantek DSO have a crap software, better get a rigol.