One of the most critical factors in astrophotography is how long your mount can track a target without manifest elongation of the stars (of course, how much elongation is too much is often a bit subjective).
With unguided tracking exposure times are bound to be shorter than if you autoguide and with an altazimuth mount they are bound to be even shorter because of field rotation.
However, being able to have at least a reasonable estimate of how long an exposure you can expect by a given combination of mount, scope and camera is very important in order to plan efficiently your photographic sessions.
I will report exclusively about my experience with the Skywatcher AZ-GTi mount and the telescopes and camera lenses I own and have used, but some general principles can be derived.
Firstly there is a rather variable aspect that can only be assessed case by case, namely the effect of the weight and length (and hence balance and torque) of any telescope on the tracking precision of the mount.
The Skywatcher AZ-GTi specs quote 5 kilograms as its maximum load capacity, and some users have even gone beyond that, up to about 6 kg, but only for visual use.
As regards astrophotography, in my experience so far about 3.5 kg with exposure times ranging from 20 to 30 seconds is the maximum I got.
It is true that in my particular case I had that result with the Bresser AR102-XS refractor, which is a bit heavier on the front (where the doublet is), hence creating some extra torque when compared, for example, with a SCT like the Celestron C5, which weighs the same but with a different weight distribution.
More in general, and it may look quite obvious, the shorter and the lighter your scope/lens/camera complex is, the better.
On the other hand a general, quantitative rule can be derived based on the resolution in arcsec/pixel allowed by your combination of scope and camera according to the well known formula:
resolution(arcsec/pixel) = 206.265 x (pixel size in μm / focal length in mm)
Binning must also be taken into account: a 2×2 binning halves the resolution.
What I have experienced being constantly true is that, unsuprisingly, exposure times (when it is possible to neglect difference in weight and balance) are inversely proportional to the resolution of your image.
So, if you can achieve 30 seconds exposures at a given resolution, if you introduce a 2×2 binning, you can consistently obtain 60 seconds exposures with the same scope/camera pair.
Also, if you get, say, 15 seconds exposure of a given quality with a setup producing a resolution of 3 arcsec/pixel, you should achieve 60 seconds exposures of the same (aesthetical) quality with a setup having a resolution of 12 arcsec/pixel.
To sum it up:
exposure time 1 : exposure time 2 = resolution 1 : resolution 2
To prevent misunderstandings, it should be clear that, if the resolution is expressed in arcsec/pixel, higher numerical values mean lower resolution, and hence the length of the exposure is inversely proportional to the resolution.