Job flow for a box job looked like this:
boxjob job1 job_type: SQLAGENT job type: step1 step2 job2 job_type: cmd
One of the runs of the box job looked like this:
boxjob starts,
job1 goes to failure after step1
job2 is still running
Application team owner for job1, restarts job1:
sendevent -E RESTARTJOB -J job1
job1 runs its normal course this time and goes to success
job2 runs its normal course and goes to success
But boxjob shows up FAILED, instead of success.
This is expected behavior because job1 originally failed and it was manually restarted by performing a RESTARTJOB on it.
Otherwise, you should see success on the boxjob, assuming everything went to SUCCESS on the first go.