Project when a disk volume will run out of space based on its current growth rate.
By sampling used disk space at two points in time, this calculator derives a linear daily growth rate and projects forward to when the volume will cross an alert threshold (e.g. 90%, matching common Nagios/Prometheus disk-alert configs) and when it will be completely full. Linear extrapolation assumes growth rate stays constant — for volumes with bursty or seasonal growth (e.g. log rotation, batch jobs), treat this as a rough runway estimate and re-sample regularly rather than a hard deadline.
Days until full
days_until_full = (total_capacity − used_now) / growth_per_day
Two points give a linear trend, which is a reasonable first estimate, but disk growth is often non-linear (bursts from log rotation, batch jobs, or backups). For production capacity planning, sample daily for at least 2-4 weeks and fit a trend line, or feed the series into a monitoring tool's forecast function (e.g. Prometheus predict_linear).
Filesystems perform worse and become harder to safely clean up as they approach full — ext4's default reserved blocks aside, fragmentation increases and some operations (like large file writes or database checkpoints) can fail unpredictably near capacity. An earlier threshold buys time to add capacity or clean up before an outage.
Negative growth (shrinking usage) means cleanup or log rotation is outpacing new data — the calculator reports 'stable or shrinking' with no full-disk ETA, since linear extrapolation of a negative trend toward zero isn't a meaningful capacity risk.