Medplum is the fasted way to build high quality, scalable and compliant healthcare applications

medplum

I've noticed if a subscription fails enough times it will try again when the server restarts? How does this work/get configured?

Hi <@1083764540237422642>  - by default, Medplum server will retry Subscription failures first after 1 second, then 2 second, then 4 seconds, etc ("exponential backoff") for up to 18 attempts (total of 72 hours after the initial change).  You can customize the maximum number of attempts using one of the Medplum extensions to the Subscription resource: https://www.medplum.com/docs/subscriptions/subscription-extensions

oh okay interesting, did see the new retry PR

and those retries do persist a server restart

Yes, the Subscription state is stored in Redis

Under the covers, Medplum uses "BullMQ" for managing async jobs.  If you are curious, you can read more here: https://docs.bullmq.io/guide/retrying-failing-jobs