BitFolk » Monitoring

Great idea. Get people to vote it up, because I'm currently working on #10 (wiki) and #25 (view/edit reverse DNS).

Reporting data transfer should be no problem, but backup space usage is a bit trickier.

It takes quite a long time (minutes) to calculate the space used because each customer is a set of subdirectories relating to the different iterations of backup, and each of those subdirectories is a large tree of hardlinks comprising the different versions of each file. Two 1MiB files will take up 2MiB if they are different, or only 1MiB if their content is identical. As a result, we only calculate your backup space usage once a day at most.

Say you had a nagios plugin checking the backup usage and it started firing. If you went and removed a bunch of files, you'd have to wait until at least the next day before your usage was recalculated. Would this be acceptable?

I agree it would be infinitely better than what we have now where you have to pro-actively check a web page to find out what your usage is.

#38 (automated warnings for exceeding backup space allowance) probably has to come first, because it's more of an issue for BitFolk that people's backups continue running even when they've used all their allowed space.

Personally I'd be happy with the backup space being upto 24 hours old.

Surely #38 could piggy back on this (or vice versa) in that the plugin reports percentage used and raises a warning at say 90%+ and a problem at 100%+.

We already have the daily usage stored in the database (that's how it's shown on https://panel.bitfolk.com/backups/), so it is just a question of how to give a nagios plugin access to that data. It's times like this that I wish I knew how to write a nice RESTful architecture, or could afford to employ someone who does. :) Bear in mind it has to be an authenticated access, also.

Backup disk space usage is now automatically reported in Nagios.