In fact, this is not a problem.
Just keep in mind that you're using a digital oscilloscope, which digitizes the signal with different sampling rates depending on the selected horizontal sweep, and then connects the digitized points with strait line while restoring the real shape of the signal.
Your first screen shows that you are measuring voltage 50 Hz with the 10 ms / div sweep and a sampling frequency of 20 kHz Ks/s
One signal period (20 ms), digitized in this mode, 20E-03 (sec) * 20E03 (1/sec) = 400 points. This is enough to properly restore and interpolate a sine wave of 50 Hz (i.e. in a period of 20 ms).
Normal display, with a sweep 10 ms / div:
Distortion of the same signal at 10 s / div sweep
Your second screen is set to sweep 10 sec / div, and sample rate on a sweep turned to 20 samples per second (20 Sa / s). I.e. one signal period 20 ms had: 20E-03 (sec) * 20 (1/sec) = 0.4 points. That means that to restore (to interpolate the points) a sine wave with less than one point in time is impossible, so you get this mess (known as "aliasing" or a false frequency) formed by the beats of the measured frequency and sampling frequency.
In order to correctly install a data collection in a digital oscilloscope one should follow a simple rule - the sampling rate must be at least 5-10 times higher than the frequency signal, in that case you will not have the issues that we just discussed.
This applies to all digital oscilloscopes and in no way connected to any particular make or model of oscilloscope or its probes.