Write tests to ensure that tracking is detached when leaving service functions

[andchiind]

Describe the new feature you would like to see
We have historically had issues with database objects being tracked in places they shouldn't, which has caused exceptions when updating the database. Currently the system is stable, but it would be easier to maintain in the future if we had tests to detect when issues occur.

Describe the solution you'd like
Any function which reads or writes to the database should clean up after itself, unless the "readonly: false" flag is set. The test itself should have access to the same context used in the functions it is testing, so we can check the tracking status of the database objects after the function is done running.

How will this feature affect the current Threat Model?
It will avoid a potential vulnerability.