React against a cancellation of a workbench job?

Hi @xtian79,

You can use the FME Server REST API to query the status of the jobs:

For if you have the job ID or:

For all jobs.

Thanks for your answer, but it is not suitable for me, because i required an automated/programmed solution, that not involves a permanent pooling for the status of one job. i will complete my question.

No need to pool foe the status if you make it part of the workflow.

Somehow you will need to get the job status to define if it is cancelled.

As far as I understand you, I can verify the status of a job from the same running workbench? is that correct?

For that you will probably need a master / slave construction, using the FMEServerJobSubmitter calling the slave workspace in the master workspace. The result of the transfomer can be used to query the job status.

Another option would be to use a default topic to notify after job completionand have a secondary proces querying for the job status. This is a more generic approach.

I think this would be a good enhancement to the FME Server System Event notifications capabilities. It might be worth submitting a request for this enhancement...

Actually, thinking a bit more about it, this is already possible.

There are default topics which get published to whenever a workspace fails on FME Server (a cancellation is considered a fail). You just need to set up an Automation (FME Server 2019+) or Subscription (FME Server <2019) to listen for the failed workspace topic notification and then process accordingly.

Sounds good, I am going to try

Hi @nic_ran, i was looking the info in the link that you posted, but that is information from 2013, I tried to look for the latest one, but there is no default topics mentioned in the current documentation.

So far, I try to create a subscription to the topic:

JOBSUBMITTER_ASYNC_JOB_FAILURE

acording to https://docs.safe.com/fme/html/FME_Server_Documentation/ReferenceManual/Monitoring_FME_Server_Activity_Using_Notifications.htm#Monitori2

I set up the subscription to the topic mentioned above, and it works perfectly when I test the topic. Funny part: the subscription does not works when i submit a job and it fails, neither when i cancel it. (I am using the FME 2018 and the workspaces in the samples folder)

Sorry for the old link. Here's the link from 2016 about default topics (and it's still current in 2019):

https://knowledge.safe.com/articles/30539/workspace-notifications-part-2-configuring-the-def-4.html

You'll note that there are different topics which get notified depending on the service you've configured your job to run under, so make sure to use the correct one (either JobSubmitter or DataDownload). If you're not using those services then you can always create your own topic/notification/subscription. This would give you more fine-tuned control as you could do it for a specific workspace.

After an implementation of a long running workspace, and be able to cancel it (I am not so fast),

I notice that JOBSUBMITTER_ASYNC_JOB_FAILURE topic is not triggered.

As seen in the image , the workspace was cancelled one and failed 3 times, and the topic monitoring show me the 3 failures activation but not the one from the one from the cancel action.

What i am missing here? is this really not supported by FME Server?