Snapshot Storage Growth Calculator

{{ summaryTitle }} {{ summaryValue }} {{ summaryLine }} {{ badge.label }}: {{ badge.value }}

Snapshot flow visual unavailable

Protected dataset:

Use logical active data, not raw pool or array capacity.

Daily change rate:

Prefer changed-block telemetry or backup deltas over net file growth.

Snapshots per day:

per day

Retention window:

days

Reduction ratio:

Use 1.0 when the workload has no measured reduction benefit.

x:1

Safety headroom:

Baseline accounting:

Current snapshot reserve:

Zero leaves this as a new sizing target.

Annual dataset growth:

Zero keeps the forecast target level.

Forecast horizon:

Choose the capacity-planning window shown in the trend.

One-day burst change:

Zero adds no custom burst case.

Metadata overhead:

Zero adds no separate metadata layer.

Clone workspace reserve:

Zero keeps clone capacity outside this reserve.

Cost per TiB-month:

Zero omits monthly cost from the reserve plan.

$/ TiB-mo

Measure	Value	Planning meaning	Copy
{{ row.label }}	{{ row.value }}	{{ row.note }}

Day	Snapshots	Retained delta	Target reserve	Gap	Copy
{{ row.day }}	{{ row.snapshots }}	{{ row.retained }}	{{ row.target }}	{{ row.gap }}

Case	Change	Target reserve	Delta vs base	Copy
{{ row.label }}	{{ row.change }}	{{ row.target }}	{{ row.delta }}

Reserve Actions

A snapshot can appear almost instantaneously because it initially references existing data, yet its storage cost grows as live blocks are rewritten or deleted. Older versions must remain available for retained restore points, so daily churn often matters more than the number of new files.

Snapshot capacity planning starts with the logical protected dataset and the share changed each day. Retention determines how many days of old blocks remain pinned. Snapshot cadence divides that daily change among restore points, but in a steady changed-block model it does not multiply the total daily churn.

Metadata, clone workspaces, deletion lag, burst churn, and safety headroom can materially increase the target. Growth forecasts add another uncertainty because they hold the churn percentage, retention, and reduction assumptions constant while the protected dataset compounds.

Snapshot capacity drivers
Driver	Effect on reserve	Frequent error
Daily change rate	More rewritten or deleted data retains more old blocks.	Substituting net dataset growth for changed-block churn.
Retention	More days keep more accumulated changes.	Counting snapshot points without considering how long blocks persist.
Reduction ratio	Compression and deduplication reduce modeled storage.	Assuming a ratio that has not been measured on this workload.
Baseline accounting	A repository may need a reduced full seed in addition to deltas.	Applying local copy-on-write behavior to every platform.

Enough capacity does not make a snapshot a complete backup strategy. Restore testing, application consistency, repository health, access separation, immutability, and an off-system copy may still be required. The reserve estimate answers how much space the stated assumptions imply, not whether recovery will succeed.

How to Use This Tool:

Build the base reserve from measured changed-block behavior, then add only the storage layers that exist in the target platform.

Enter the logical Protected dataset and choose TB, TiB, PB, or PiB. Do not use raw array capacity.
Set Daily change rate from array telemetry, changed-block reports, or backup deltas, then enter Snapshots per day and Retention window.
Enter a measured Reduction ratio, or use 1.0 when no compression or deduplication benefit is known. Add Safety headroom after the modeled layers.
Choose Baseline accounting. Use changed blocks only for a delta reserve; include one reduced seed copy only when the repository stores or bills it.
Add current reserve, growth, burst, metadata, clone, or cost assumptions only when they apply. Compare the reserve plan with stress cases and the forecast before changing capacity or retention policy.

Interpreting Results:

The reserve target includes retained changed blocks, any seed copy, metadata allowance, clone workspace, and headroom. Sizing target means no current reserve was entered. Reserve covered means current reserve is at least the modeled target, not that every recovery requirement is satisfied.

A positive current gap is a shortfall. When the gap is more than 20% of the target the state becomes Capacity shortfall; a smaller positive gap is Reserve tight. Check the high-churn and burst cases before treating a narrow margin as sufficient.

Technical Details:

All capacity values are normalized to tebibytes (TiB). Decimal TB and PB are converted by their byte definitions, while 1 PiB equals 1,024 TiB. Percentages are applied to logical dataset size before reduction.

Formula Core:

The base model accumulates daily logical churn across retention, divides the retained data by the reduction ratio, adds optional storage layers, then applies headroom.

\begin{array}{rcl} L & = & D \times c \times d \\ R & = & \frac{L}{q} \\ S & = & R + Seed + Metadata + Clone \\ Target & = & S \times (1 + h) \end{array}

D is the dataset in TiB, c is daily change as a decimal, d is retention days, q is the reduction ratio, and h is headroom as a decimal. A seed is D ÷ q when enabled. Metadata is a percentage of reduced retained changes, while clone reserve is a percentage of the dataset divided by the same reduction ratio.

Snapshot count equals snapshots per day multiplied by retention days. Average logical change per snapshot equals daily logical churn divided by snapshots per day, so cadence changes point count and per-point size but not steady-state retained churn in this model.

Mechanism Core: Stress and Forecast Rules

Snapshot stress and forecast rules
Case	Rule	Interpretation
Quiet	50% of entered daily churn	Lower-change comparison.
Base	Entered daily churn	Current planning assumption.
High churn	200% of entered daily churn	Persistent high-change comparison.
Burst day	One normal retention day is replaced by the entered burst rate	Single-day spike, not a sustained rate.
Growth forecast	Dataset × (1 + annual growth)^months/12	Churn percentage and policy remain fixed.

Worked Check:

A 12 TiB dataset changing 2.5% per day retains 9 TiB of logical changes over 30 days. At a 1.4:1 reduction ratio that becomes about 6.43 TiB. Adding 20% headroom produces a reserve target of about 7.71 TiB. Four snapshots per day create 120 restore points, but do not change that base target.

Limitations:

This is a planning model, not a platform inventory. Real consumption can differ because of block size, write amplification, deletion timing, metadata layout, sparse files, database behavior, encryption, compression order, deduplication scope, replication, consistency groups, and provider billing rules. Replace assumptions with observed data and verify the result against the storage platform.

References:

Definitions of the SI Units: The Binary Prefixes, National Institute of Standards and Technology.
zfs.8, OpenZFS documentation.
How to create and manage Elasticsearch snapshots, Simplified Guide.