Taming journald on a small VPS

2026-09-14 · Alex Moreno

A small VPS I use for side projects started warning about disk space last month. The culprit was not a runaway database or a forgotten backup — it was /var/log/journal, sitting at a little over 4 GB.

Finding out what is actually there

journalctl --disk-usage
journalctl --list-boots | tail

The journal had been keeping every boot since the machine was provisioned. Most of the volume came from one chatty service that logged every health check.

The settings that mattered

In /etc/systemd/journald.conf I ended up with just three lines:

SystemMaxUse=300M
MaxRetentionSec=2week
RateLimitBurst=2000

Then systemctl restart systemd-journald and a one-off journalctl --vacuum-size=300M to reclaim the space immediately.

What I would do differently

Set these on day one. Defaults are tuned for machines with far more disk than a five-dollar instance, and the journal will happily grow until something else breaks.

← All posts