When will your logs fill the disk?

Unrotated logs grow quietly until the day they don't fit — and a full disk takes the whole app down with it. Enter your log volume and disk, see the fill date, and copy a logrotate config that caps it for good.

The outage that grows quietly

Disk-fill is one of the most boring and most common ways a backend dies. Nothing is wrong — until the log file that's been growing a little every day finally doesn't fit, writes start failing, and the database, the web server, or the whole box falls over at once. It's preventable with one config file, which is exactly why it's so frustrating when it happens.

Setting up logrotate once is easy. The harder part is knowing it's actually working on every host, that a noisy new service didn't blow past its cap, and that you'll hear about a filling disk before it's full rather than after. That continuous watch — on every box you own — is the kind of thing a control plane keeps an eye on for you.