NAPPS-2280: fix deletion timeout

[Kat-Alo]

On dev and prod, deleting pages with lots of snapshots results in a 504 timeout error. My suspicion is that it's because the Rails deletion of records with children is a very costly operation.

Luckily, folks at an org called Census had a very similar problem and created a little tool they call Miss Hannigan. The tool nullifies the relationship between the children snapshots and the parent page and purges the orphaned children records ansynchronously.

If we are concerned about relying on this particular (seemingly little-known) gem, we could port some of the logic into Klaxon — not sure what that protocol looks like.

To test

• docker-compose down && docker-compose build && docker-compose up
• sign at http://localhost:3001/klaxon/ (you won't receive a sign-in email since you're running things locally. Look for the HTML of the body of the email in your terminal to find the link w the sign-in token)
• Add a page that will change frequently, so we know it'll generate a snapshot
• Wait for a snapshot to be generated
• in a new terminal...
  • klaxon-psql
  • \c klaxon;
  • select page_id from page_snapshots; (to confirm a snapshot has been generated for a page)
• Back in the UI, delete the page
• Return to the terminal window with the psql shell and confirm the snapshot you saw is no longer there

The real test will be getting these changes to dev and confirming we can delete the big record that's been causing a timeout. Then, we'll be able to open a PR with the original Klaxon.

[Kat-Alo]

Just a note that I'm not sure what these tests are given that I didn't think we had tests in our workflow for Klaxon yet?? 🤔

[Kat-Alo]

Also for posterity: I opened an issue with the original Klaxon.

[jtotoole]