TODO.md

TODO

Changes

• make all trigger function names follow the format [before/after]-[insert/update]-[table_name]-tg-fn

• make all trigger names follow the format [before/after]-[insert/update]-[table_name]-tg

• extract all functions currently in tg functions into their own functions (for resusability)

  • e.g: process-jobs is good

• update all trigger functions to be wrappers that use other named functions (for readability and resuability)

• ensure there are unique constraints on the job and worker columns of the current_job table

• test out using the 'skip locked' line in the process jobs functions

• update the vault secret from supabase_serevice_role key to supaqueue_secret

  • in the vault
  • in the pg function that calls the edge function
  • in the edge funciton itself (must be added to env variables)

• add an 'updated_at' column and tg to the jobs table, default now()s

  • sort the jobs in the process_jobs table on updated_at
  • add attempts filtering to the processing jobs fn.

• support HTTP Methods (pass payload as body of req)

  • POST
  • GET (currently payload gets passed as query params, not body)
  • PATCH
  • DELETE

• support explicit body and query params

  • should I make different paths? or use if/else in one function call?

• UPDATE THE migration.sql file with changes

• update the readme.md file with changes.

Testing

• multiple concurrent queues
• long running api calls
• adding a job to the queue while queue is being processed
• adding an API secret to the queue
• calling a real edge function.

Feature Ideas

• timebase rate limiting per queue
• capturing the results of a job in the job table
• capturing an array of errors from a job in the job table
• add cron to 'kick' the queue every minute.
• adding in crons to clean up jobs table every [day/week/month]
• host the edge function in a node package, and make it importable to a custom edge function, scuh that users don't need to copy and paste code, just update version # for updated edge function.