We have jobs which are not run every week

[theseal]

If the cleanup job removes output files from jobs run less then every week we get alert from the monitor.

[mickenordin]

LGTM

[mikaelfrykholm]

This looks scary and prone to errors.
# m h dom mon dow command
@yearly /usr/local/bin/scriptherder apa

❯ ./cleanup_scriptherder
 Traceback (most recent call last):
   File "/tmp/./cleanup_scriptherder", line 63, in <module>
     job_name = tokens[position].replace("-", "_")
                ~~~~~~^^^^^^^^^^
 IndexError: list index out of range

[theseal]

This looks scary and prone to errors.
I would agree if we were editing the crontab by our self but in this case it's puppet doing all the magic, consistency.

# m h dom mon dow command @yearly /usr/local/bin/scriptherder apa

❯ ./cleanup_scriptherder
 Traceback (most recent call last):
   File "/tmp/./cleanup_scriptherder", line 63, in <module>
     job_name = tokens[position].replace("-", "_")
                ~~~~~~^^^^^^^^^^
 IndexError: list index out of range

Can try to add some error handling so we don't crash but the only "bad" thing if this script doesn't match a requested scriptherder job is that might get a disk event after a while.

[mikaelfrykholm]

LGTM