NINA's autofocus routine, if you have tilt or significant backfocus error, can sometimes give incorrect results as it measures star size across the whole field. Instead of a nice tight curve with relative small error bars:

You might instead see a flat bottom with large error bars where it is unable to determine the precise smallest hfr position since it varies largely across the frame. Using the ROI setting in the AF settings can help clean this up as it excludes off axis sections of the field.

Collimating the primary is actually a lot easier than the secondary in my experience.
While secondary collimation involves a lot of sticking your whole hand onto the mirror and trying to hold it in place while simultaneous staring down an eyepiece and tightening the right screws (and if you're me, not dropping your tools into the tube); primary collimation is simply about aligning the reflection of the focusing cap (black circle outlined in red) to the primary mirror mark (black circle outlined in green). The white spacing (with purple lines) should be equal around the donut.

The blue donut is the secondary optical center mark, it being misaligned is probably why your illumination is slightly off center. In your case the primary mark isn't perfectly positioned also.
I would collimate the primary by mounting the scope on my EQ mount, slewing to zenith during the daytime (obviously when the sun wasn't shining on the scope), and loosening all the push and pull screws (and the side screws further securing the mirror cell) to a point where the mirror cell would have a lot of give (about 1 cm). Then I would go around and around slowly tightening each of the three larger screw while keeping the position of the two circles in alignment - once each screw was about as tight as it could get I'd snug up the 6 smaller screws.