Under unknown circumstances it is possible for weekly schedules to fail to be generated after sending (and/or failing to send) a push notification.
We only have one known participant with this issue, this participant was an iOS test participant on our staging server. The cause of this is unknown, but presumably there is an unhandled error in the push notification logic - weekly notifications trigger their own generation of the next event, this seems brittle, we should change that.
1) Make a script that checks whether there are missing weekly ScheduledEvents for participants and run it on deployments to determine extent of the problem.
2) Make a periodic task that checks existing weekly ScheduledEvent, clears out ANCIENT (2 week old?) weekly ScheduledEvents, and populates any missing current or next-week++ weekly ScheduledEvents.
3) Junk the current brittle implementation.
Under unknown circumstances it is possible for weekly schedules to fail to be generated after sending (and/or failing to send) a push notification.
We only have one known participant with this issue, this participant was an iOS test participant on our staging server. The cause of this is unknown, but presumably there is an unhandled error in the push notification logic - weekly notifications trigger their own generation of the next event, this seems brittle, we should change that.
1) Make a script that checks whether there are missing weekly ScheduledEvents for participants and run it on deployments to determine extent of the problem. 2) Make a periodic task that checks existing weekly ScheduledEvent, clears out ANCIENT (2 week old?) weekly ScheduledEvents, and populates any missing current or next-week++ weekly ScheduledEvents. 3) Junk the current brittle implementation.