Anyone can make a backup of any site by retrieving the system/export.json from the site.
The export.json file differs from the format pages are routinely saved. The wiki-server implementation prefers to keep pages in separate files within a pages/ directory.
An administrator with ssh access to a server can reconstruct the pages directory from an export file using the json utility 'jq'. stackoverflow
cat export.json | jq -c -r 'keys as $slug|"\($slug)\n\(.[$slug])"' | while read -r slug ; do read -r item printf "%s" "$item" | jq . > "pages/$slug" done
Here is a slower but slightly simpler version of this script.
for slug in `cat export.json | jq -r 'keys'` ; do cat export.json | jq '.["'$slug'"]' > pages/$slug done