I'm a GIS analyst who does spatial analysis, makes maps (static and web-based), and I have an undergraduate background that taught us some of the history of map making.

To answer your first sub-question, no, north was not always "up". Even maps that are made today do not necessarily have north at the top, although it is rare that a map of the globe would be oriented without north at the top. In antiquity there weren't any true "global" maps because the ecumene of the known world was limited. The scale of the map and what the map was depicting often influenced its orientation. Major cities or water bodies were commonly oriented near the top in relation to whatever else was being mapped. Here is a neat website that allows you to pan around the globe, zoom in and out, and filter the date range to find older maps:

https://www.oldmapsonline.org/map/cuni/808729

The following link shows the orientation of one of the older maps from the middle east that I was able to find: http://cuni.georeferencer.com/compare#map/804823034252

You can view the details of the map as a document by clicking on "view this map" in the following link: https://www.oldmapsonline.org/map/cuni/808729

In this example the cartographer oriented the map so that the Mediterranean/Levantine sea is at the top.

Orientations changed all the time depending on what major geographic features were being depicted. Map makers today will still often change the orientation of large-scale (small area) maps if it makes sense/makes thinks look nicer given what is contained within the map extent.

In relation to global maps always being "up", some speculate that this is a function of a European-centric world-view, but there's really only 2 options for what "should" be "up" on a world map: the north pole or the south pole. Mathematically the globe is divided into meridians that stem from the poles (which are a logical place to divide the world from) and parallels at equal distances from the equator (again, a logical place to start from mathematically/geometrically). Two early mathematicians/geographers named Hipparchus and Ptolemy originate this concept in the first/second centuries, and Ptolemies map depicted north at the top. Here is a link to a journal article on the history of maps that was published in 1879 that seems to be freely accessible:

https://www.jstor.org/stable/196503?seq=1#metadata_info_tab_contents

It shows versions of both Hipparchus' and Ptolemy's maps with north at the top. Again, mathematically/logically either north or south goes at the top, and these early maps happen to depict north at the top. It's likely that these early world maps influenced later cartographers who illustrated global-scale maps.