As far as I am aware, absolute uncertainty means the absolute amount by which the measured value may differ from the actual value. This is as opposed to fractional uncertainty, which is absolute uncertainty divided by the measured value.
And I find "systematic uncertainty" conceptually awkward. Systematic error is the more usual expression.
So I would say that total error is systematic + random, where each of those may be (consistently) interpreted as fractional or absolute.
That is with regard to repeated measurements which are in principle of the same quantity. If they are for different quantities (because some parameter is being varied) these errors may vary in different ways. E.g. the systematic fractional error my remain constant, while for random error it is the absolute error that is constant.