Forecast future storage needs from historical growth rate and planning horizon.
Given a constant daily growth rate, the runway before a disk fills is simply the remaining free capacity divided by that daily growth — the same math behind Kubernetes and cloud-provider disk-full predictions. This linear forecast assumes growth stays roughly steady; sudden traffic spikes, log storms, or new workloads landing on the volume can shrink the runway far faster than the trend suggests, so this estimate should be revisited regularly rather than treated as a one-time calculation.
Days until full
days_until_full = (total − used) / daily_growth
Take used-space samples (df -h or equivalent) at the same time each day over at least a week, then average the day-over-day differences — a single day's sample can be skewed by one-off events like a large backup or temp file cleanup.
Set a monitoring alert well before the projected date (e.g. at 80% capacity), and start capacity expansion, cleanup, or data archival work with enough lead time to complete before urgency escalates — waiting until the disk is nearly full risks application failures or an emergency, unplanned expansion.
Growth rates are rarely perfectly linear — traffic patterns, seasonal spikes, new features generating more data, or scheduled cleanup jobs (log rotation, cache eviction) can all accelerate or decelerate the trend, so this figure is best used as an early-warning estimate rather than an exact prediction.