Resume Send Event Task

[dcwik96]

How can I make Send Event Task a deadletter job to be able to resume it? I made it Asynchronous. After that task, I have outgoing conditional sequence flow where I check if variable is set properly. When FlowableException is thrown there is also org.springframework.amqp.rabbit.support.ListenerExecutionFailedException.
I assume that there is problem with transaction so I put Send Event Task in subprocess which is transactional.

What else can I do to be able to resume Send Event Task using REST API (deadletter or maybe something else?) ?
How to deal with transaction of Send Event Task?

[amporsim]

Hi @dcwik96

What else can I do to be able to resume Send Event Task using REST API (deadletter or maybe something else?) ?

Please have a look to a previous forum post:

Deadletter Jobs vs REST API

Hi, Is there any existing REST API to monitor / resume Deadletter Jobs, or any plan to build one? Note I have not found such API reading the documentation. Currently, I use a custom resource to do that. Would you accept a PR for that?

Best regards,
Simon

[dcwik96]

How can I make SendEventTask as deadletter? There is an Flowable Exception thrown if something goes wrong

[SIVA_EXPERT]

@dcwik96, I found this in the previous discussions which might help.

Dead Letter Jobs Retry

Hello everyone, I have a question regarding the jobs in the dead letter table and maybe the event subscription table: I am building a REST controller to view all jobs that resulted in an error. Is it possible to set the process variables of these jobs after they run into the error and move them back to executable? I tried already using the runtimeService to set the variables but it did not have any impact. I know that moving jobs from dead letter table to executable table could fire a retry, b…

[joram]

@dcwik96

How can I make SendEventTask as deadletter? There is an Flowable Exception thrown if something goes wrong

Yes, and that’s correct. After a configurable number of retries (3 by default), the job is moved to the deadletter jobs. From there, it can be moved back again to be executed again.