The backup scripts should do SOMETHING which makes it easy for monitoring systems to evaluate their state. I would propose generating a text file at the end of each run, the contents of which looks like this:
last_backup: <unixtimestamp>
backup_result: [success|failure]
backup_bytes_WHICH: <bytecount>
This should just be dumped into the filesystem somewhere; a monitoring system can look for it, e.g. by polling a static web server. The monitoring system should at the very least alert if backup_result isn't "success" (an easy regexp even w/o a parsing), while fancier implementations could parse the file and graph the bytes etc.
The backup scripts should do SOMETHING which makes it easy for monitoring systems to evaluate their state. I would propose generating a text file at the end of each run, the contents of which looks like this:
This should just be dumped into the filesystem somewhere; a monitoring system can look for it, e.g. by polling a static web server. The monitoring system should at the very least alert if backup_result isn't "success" (an easy regexp even w/o a parsing), while fancier implementations could parse the file and graph the bytes etc.