A job submit by the FMEServerJobSubmitter will never post to a topic on success or failure, even if it was registered with it at publishing time. Nor is there a way to set a topic to post to within the FMEServerJobSubmitter. It would be great to be able to chain and control jobs with the FMEServerJobSubmitter, and still make use of the power of FME Server notifications/automations.
Open
FMEServerJobSubmitter Post to Topic
+9- Safer
There do exist a few ideas for workarounds at this question. However, full compatibility of topics with the FMEServerJobSubmitter would be an intuitive enhancement.
+19
I've tried some of the workarounds but none of them are really great work with. Especially if you are interested in the job id you have to extract it out of JSON or if you want to start multiple jobs in parallel or on a "child" engine is tricky or not supported.
So please add Topic-Support in FMEServerJobSubmitter to build automated processing pipelines with notification support because the automations can't replace a complex FME-Control-Workspace :-)
+1Can I get a million up votes on this one??? This would be a game changer especially for job that were run in parallel and not waiting for the jobs to complete. Currently there is no way to handle a failed job when all jobs are set to "not wait to complete". A user has to manually monitor the jobs failed status and handle manually. Which is quite tedious to go through 100 jobs 1 by 1 just to get the parameters that were used in execution to kick them off again.
At a very minimum the "JobSumitter" transformer should notify the JOBSUBMITTER_ASYNC_JOB_FAILED and JOBSUBMITTER_ASYNC_JOB_SUCCESS out of the box topics. Ideally it should expose the same properties when you publish a workspace to server.
- Send HTTP Message Body to Reader
- Post to Topics on Success
- Post to Topics on Failure
- Post Data from Writer
The only reason I am using the job submitter in the first place is because you can specify a "max run time" of a job. If FME server had that parameter in the scheduler then I wouldn't need to to user the JobSubmitter transformer
+15- Contributor
An addition to this idea (had this need today) would be to add a option on the FMEServerJobSubmitter (like the Automation workspace runner) where you get a number of retries on fail.