Skip to main content

Hey all,


Happy Sunday!


I’ve written my code using the @mondaycom/apps-sdk Queue. In the development environment, there is no queue implementation; it directly makes an HTTP call to the job endpoint.


I’ve made some assumptions about retry and back-off behaviour while developing the code. However, I couldn’t find detailed information about the Queue in the documentation. Could you please help clarify the standard behaviour of the queue, especially regarding the number of retries in case of failure?


Additionally, is there an endpoint to query the number of failed jobs or status of a message enqueued? What is the back-off policy, and can we set it based on the type of error? Currently, I’m assuming it’s not customisable, so for permanent failures, I return a 200 status and send a notification to the user.


Are there any other important details I should consider?


I’m feeling a bit edgy to go with this implementation as I’m approaching app submission to the marketplace soon. Your guidance would be greatly appreciated.


Thank you for your help!

Sourav

@gregra do you think you could answer these questions?

Thanks, David!


Found in the documentation that the queue makes 9 retry attempts before finally failing.


Screenshot 2024-07-03 at 3.24.47 AM


@gregra can you please confirm this? Also, what backoff policy is applied with the retries? My code depends on a third-party API, and there might be times when the upstream service is unresponsive, so an incremental backoff delay strategy would be ideal.


Additionally, what are the recommended best practices for handling permanent failures of a background job? Currently, I send a notification to the user.


Thanks! 🙂

Hello there @souravray !


I will be checking all of this with our team and get back to you 😀

Hello again @souravray,


I heard back from the team!


There are 9 retries, it cannot be set by the type of the error and it’s not customizable.


The retry will kick in if you return error or you didn’t answer with any response after 10 minutes.


Regarding the question about endpoint to query the failed jobs or status of a message enqueued, that is currently not supported.


Regarding the other questions, we will work on documentation for it and we will send it here once it is ready 😀


Cheers,

Matias

@Matias.Monday Would it be possible to add the queue incremental backoff delay strategy as a feature request? I think it’s very common to use queues for interfacing with APIs, so this type of backoff policy would be very helpful for many integrations.


Without this feature, it really limits the usefulness of queues in monday code.

Hello there @PluginGenie!


You can open a request here 😀

Reply


Terms & ConditionsCookie settings