fix(medusa): add support for retrying failed event bus jobs #2566
Conversation
🦋 Changeset detectedLatest commit: 71d6a11 The changes in this PR will be included in the next version bump. This PR includes changesets to release 1 package
Not sure what this means? Click here to learn what changesets are. Click here if you're a maintainer who wants to add another changeset to this PR |
| @@ -192,9 +201,15 @@ export default class EventBusService { | |||
| }) | |||
| return await stagedJobRepository.save(stagedJobInstance) | |||
| } else { | |||
| const opts: { removeOnComplete: boolean; delay?: number } = { | |||
| const opts: { removeOnComplete: boolean } & EmitOptions = { | |||
There was a problem hiding this comment.
todo/suggestion(if-minor): Move removeOnComplete to EmitOptions and default value to options parameter instead of here
|
Hmmm, I'm not sure this is the best for most of our use-cases with subscribers. I often have 5-6 subscribers on a single emit type (ex: order.placed). If any of the individual subscribers fail, based on the implementation I'm seeing here, they will all get re-run correct? So all the individual subscribers need to be smart enough to ignore retries if another subscriber failed. Ex: If I send an email to the customer on order placed, but also have another subscriber (maybe from a plugin I installed) that continues to fail, the customer could receive lots of emails depending on the retry-policy, unless there is business logic that's checking to see if the email has already been sent? In the past I've used medusa-extender to override the event-bus and add retries using try/catch in the worker instead of doing it the Bull way. That way I was only retrying the failed events and not spamming successful subscribers. Just my 2 cents! |
Closes CORE-770