Hopefully preventing server crashes from running out of memory

[misaugstad]

(hopefully) fixes #3224

This PR improves the performance of two of our biggest API calls: /adminapi/labels/cvMetadata and /attributesWithLabels. Instead of trying to convert all of the labels to JSON in one go like we had before, which forces us to load everything into memory, we are now writing 10k labels at a time to JSON files. We can then transfer the JSON files with much less memory.

I tested this out on my dev environment by artificially lowering the amount of memory available to Java so that I could test with smaller datasets. This seems to be solving the issue. Will only know for sure once it is on production.

Additionally, I only did this with two API endpoints to start. If we determine that the method is working, it's pretty easy to add that workflow to other endpoints!

In the saved file we include the timestamp for when it was created, which will allow us to serve multiple concurrent API requests for the same API.

Things to check before submitting the PR

• I've written a descriptive PR title.
• I've added/updated comments for large or confusing blocks of code.

[misaugstad]

Slight correction: I did this for a third API endpoint too, the /attributes API!