Closed sfisher closed 1 year ago
Polling endpoints are defined in the "experimental" section here: https://petstore.swagger.io/?url=https://raw.githubusercontent.com/CDLUC3/mrt-dashboard/main/swagger.yml#/experimental
Documentation for the callback feature can be found here: https://confluence.ucop.edu/display/UC3/Merritt+Callback Note that the callback is configured on a per-collection basis.
After discussion with the team it sounds like it's preferred to poll to avoid some difficulties having callbacks work with non-public development servers. Ryan would like more than once a minute if possible. Ashley mentioned that she doesn't love crons and that a daemon might be preferable for polling, also.
Terry mentioned he believes that these would work with both arks and dois (local ids). The experimental ones under https://petstore.swagger.io/?url=https://raw.githubusercontent.com/CDLUC3/mrt-dashboard/main/swagger.yml#/experimental
This set of tasks is complete and tested by me and Ryan. However I moved the last task to a separate ticket since we'll need to work with Ashley to add to systemd and for production it will need to startup and have cron removed around the time of the deploy.
Merritt would like us to get submission status information from API URLs instread of using the notifier and polling the OAI-PMH feed.
We would want to poll hitting their API endpoint (or use their callback) to get status information for outstanding items.
RAILS_ENV=development rails merritt_status:update
. pid and logs are is in our logs directorylog/merritt_status_updater.pid
. It can be stopped by reading the pid andkill -15 `cat my/path/to/log/merritt_status_updater.pid`