When Auto-Wrangling is enabled, it can detect a bad job, automatically block it, and notify the user via email.
In certain situations, due to timing issues, the system may generate several email messages before it finishes blocking the job, which can become a nuisance for some.
This is especially the case when the job in question uses the "cmdrange" jobtype backend, since, by default, it will move onto the next agenda item even if the current one fails.
To minimize such behavior, jobs can be told to migrate their instances when an agenda item fails by setting its "migrate_on_frame_retry" flag.
To have 'migrate_on_frame_retry' set on every single job submitted, add it to the definition of supervisor_job_flags in the supervisor's qb.conf file.
http://docs.pipelinefx.com/display/QUBE/supervisor_job_flags
You should be able to confirm that the flag is set on a job by checking the "flags" field in the output from "qbjobs -l <JOBID>" or the job details in the GUI.