### Description
Despite making sure only a small range was used for each value, this test can nonetheless result in an explosion of memory usage: it has twice in a row resulted in my poor laptop filling its swap in a matter of seconds, forcing me to kill the test or hard-reboot my machine.
This small PR lowers the high bound of the range used for the generator, to a value that seems to consistently allows the test to finish in a more constrained environment.
Additionally, it also sorts the tests alphabetically: it makes it much easier when scrolling through the output to find a specific test.
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3942
GitOrigin-RevId: c7a786511fe82701fab29cf164dc3fbbe77f4262
I discovered and removed instances of Boolean Blindness about whether json numbers should be stringified or not.
Although quite far-reaching, this is a completely mechanical change and should have no observable impact outside the server code.
PR-URL: https://github.com/hasura/graphql-engine-mono/pull/3763
GitOrigin-RevId: c588891afd8a6923a135c736f6581a43a2eddbc7
