I've got about a dozen servers that each have crontabs with anywhere from 20-50 crontab entries. My single most common cause of a process failure is someone commenting out jobs in cron during a fix or patch and then forgetting to uncomment the jobs.
I'd like to do two things to solve this:
- Start using our schedule suppression process that allows users to suppress schedules without actually touching crontab. Nothing magical - just touch a file in a directory dedicated to the process. The process checks that directory on start-up.
- Implement a process that will send out alerts if crontab doesn't match its backup or current version in svn.
Can anyone recommend an existing solution for #2 (alert when crontab changes)?