monit pool start brings back only dead nodes, keeps alive working

log.txt

~/d/project (feature/delayed_job_pool|✚3…) $ cat tmp/pids/delayed_job.0.pid 
14702
~/d/project (feature/delayed_job_pool|✚3…) $ cat tmp/pids/delayed_job.1.pid 
14704
~/d/project (feature/delayed_job_pool|✚3…) $ ps aux | grep delayed
rupert           14702  72.9  2.9  4865820 481384   ??  R     9:43AM   3:25.55 delayed_job.0    
rupert           14704   0.0  2.8  4839192 467044   ??  S     9:43AM   3:26.31 delayed_job.1    
rupert           15205   0.0  0.0  4267752    360 s005  R+   10:18AM   0:00.00 grep --color=auto delayed

~/d/project (feature/delayed_job_pool|✚3…) $ kill -9 14702
~/d/project (feature/delayed_job_pool|✚3…) $ ps aux | grep delayed
rupert           15431   0.0  0.0  4267752    688 s005  R+   10:18AM   0:00.00 grep --color=auto delayed
rupert           14704   0.0  2.8  4839192 467244   ??  S     9:43AM   3:27.63 delayed_job.1    

~/d/project (feature/delayed_job_pool|✚3…) $ script/delayed_job --pool=high_priority,default --pool=high_priority,default,low_priority start
pid-file for killed process 14702 found (/Users/rupert/develop/project/tmp/pids/delayed_job.0.pid), deleting.
ERROR: there is already one or more instance(s) of the program running

~/d/project (feature/delayed_job_pool|✚3…) $ ps aux | grep delayed
rupert           14704   0.0  2.8  4840216 468344   ??  S     9:43AM   3:30.39 delayed_job.1    
rupert           15556   0.0  0.0  4258468    188 s005  U+   10:19AM   0:00.00 grep --color=auto delayed
rupert           15522   0.0  1.0  4502284 172380   ??  S    10:18AM   0:01.93 delayed_job.0  



~/d/project (feature/delayed_job_pool|✚3…) $ kill -9 14704
~/d/project (feature/delayed_job_pool|✚3…) $ ps aux | grep delayed
rupert           15623   0.0  0.0  4267752    628 s005  R+   10:21AM   0:00.00 grep --color=auto delayed
rupert           15522   0.0  1.5  4596496 248832   ??  S    10:18AM   0:15.24 delayed_job.0    

~/d/project (feature/delayed_job_pool|✚3…) $ script/delayed_job --pool=high_priority,default --pool=high_priority,default,low_priority start
ERROR: there is already one or more instance(s) of the program running
pid-file for killed process 14704 found (/Users/rupert/develop/project/tmp/pids/delayed_job.1.pid), deleting.

~/d/project (feature/delayed_job_pool|✚3…) $ ps aux | grep delayed
rupert           15658   1.9  1.0  4496152 169664   ??  S    10:21AM   0:00.87 delayed_job.1    
rupert           15522   0.1  1.5  4601616 256520   ??  S    10:18AM   0:17.28 delayed_job.0    
rupert           15693   0.0  0.0  4276968    892 s005  S+   10:21AM   0:00.00 grep --color=auto delayed