How to handle command errors about cqrs

[sh3d2]

Hi,

how about changing ICommandHandler.execute
from

    execute(command: T, resolve: (value?) => void);

to

    execute(command: T): Promise<any>;

that way we'd have full information about whether command handling was successful or not,
and handlers wouldnt have to deal with that ugly-looking-callback-style resolve 😄

[moltar]

Using CommandResult would not make it possible to retry failed jobs though that are managed by some queue system. E.g. RabbitMQ. If a job failed, it'd go back to the queue and be retried again. This is a very useful pattern. But if all jobs always succeed, then this pattern is not possible.

[pawpartyka]

+1

[kamilmysliwiec]

Hi,
The only way is to use resolve() method. Something like CommandResult sounds fine. 🙂

[mvanderkroon]

@moltar I am looking exactly for the behaviour you mention: jobs managed by some queue system, such that in case the handler fails either caused by a (transient) exception, or the instance is shutdown, the job is retried automatically. Could you point me towards any resources/examples (anything really) on how to configure NestJS+Sagas in a manner to achieve exactly that?