LearnHealth & Safety
How To Know Storage Is Running Out Before It Does
1 min readUpdated 6 Nov 2026
A full disk is never a surprise in hindsight. The growth was visible for months. The trick is watching the trend instead of the free-space number.
Watch the trend, not the free number
Free space tells you where you are. Growth rate tells you when you run out. A database that adds 40 GB a month on a drive with 500 GB free has a date, and you can name it.
- Growth per week per database, separated from other databases on the same drive.
- Transaction log growth, which can be far faster than data growth.
- tempdb usage, which spikes with the busiest job rather than with the calendar.
Autogrowth is a warning, not a plan
Growth events are expensive and they happen at the worst moment — during a load, while everyone is watching. If growth events show up during business hours, the file sizing is already behind reality.
- Growth in small fixed amounts causes frequent, painful events.
- Growth as a percentage grows with the file, which is usually too slow for large databases.
- Log files that never shrink and never get sized ahead of a bulk operation.
The three numbers to keep
- 1.Days until each drive fills at the current rate.
- 2.Days until each log file hits its limit or its drive.
- 3.The largest single growth event in the last month, so you know the worst day.
Storage problems are purchased in advance. A forecast that says 'eighty days' turns an outage into a ticket.
What to remember
- Forecast from growth rate, not from free space.
- Track data, log and tempdb separately.
- Growth during business hours means sizing is behind.
- Keep days-to-full visible for every drive.
