Kryocraft

Website backup storage and restore time calculator

Having a backup is the easy part. Enter your site size, how often it changes and how fast your connection is, to see how much storage the copies need, whether uploads keep up, how much work you could lose and how long a restore would take.

Start from

Your site

Files plus database

Share of the data that changes

Text compresses, photos barely

Schedule and retention

1 to 5

Connection and price

Backups rarely get all of it

DNS, checks, import, testing

0 hides the cost

One complete copy
15.0 GB
Most you could lose
6.0 hours
If one backup fails
12.0 hours
Storage saved by changes-only
70%

Complete copy every time

Stored (one copy)
570 GB
Across 2 copies
1.14 TB
Each scheduled run
15.0 GB · 57 minutes
Worst-case restore
59 minutes
Storage per month
$11.40
Pieces to apply
1

Weekly complete copy, changes in between

Smaller
Stored (one copy)
169 GB
Across 2 copies
338 GB
Each scheduled run
150 MB · under a minute
Worst-case restore
63 minutes
Storage per month
$3.38
Pieces to apply
28

Everything here is an estimate from the figures you entered. Restore time is the data coming back at your download speed plus the overhead you set, and it does not include waiting on a provider. The only real test of a backup is restoring it.

What this backup planner does

Most advice about backups stops at "make sure you have one". That is the easy half. The harder questions are practical ones. How much storage will the copies take up? Can tonight's backup finish before tomorrow's starts? If the site disappeared this afternoon, how many hours of work would be gone, and how many hours would it take to get everything back?

This tool answers those from numbers you already know or can look up in a minute: how big the site is, how much of it changes in a day, how often a backup runs, how long you want to keep copies, and how fast your connection is. It compares two common ways of storing backups side by side, so you can see what each one costs in space, upload time and restore time.

It does not make backups for you, connect to your hosting account or store anything you enter. It is a planning tool for deciding what a sensible backup arrangement looks like before you pay for one or rely on one.

The two numbers that matter

Every backup plan comes down to two figures, and they are worth knowing by name.

Recovery point is how much recent work you can afford to lose. If a backup runs once a day, a failure just before the next run loses almost a full day of changes: orders, comments, uploads, edits. If it runs every hour, you lose at most an hour. The tool shows this as "Most you could lose".

Recovery time is how long the site is down while you restore. It depends on how much data must come back, how fast it arrives, and how much work surrounds the transfer. The tool shows this as the worst-case restore.

These two pull in different directions. Running backups more often shrinks the first number and usually stretches the second, because a restore may have to apply many small pieces in order. Keeping fewer pieces speeds a restore up but puts more data at risk. The comparison cards exist to show where that balance falls for your site.

How the storage figure is worked out

Everything starts with the size of one complete copy. The tool takes the live size of your files and database and reduces it by the compression saving you enter. Text, code and database content compress well, often by half or more. Photos, video and already-compressed archives barely shrink at all, which is why a media-heavy site should enter a much smaller saving than a text-heavy one.

From there it works out how much data one backup contains and how many backups you want to keep. The retention fields work in three layers:

  • Days of frequent backups. The window in which you can go back to any scheduled run.
  • Extra weekly copies. One older copy per week beyond that window.
  • Extra monthly copies. One still older copy per month, for the mistakes nobody notices for weeks.

Layering is how people keep a long history without paying for a copy of every single run. A site might keep a week of frequent backups, a month of weekly ones and half a year of monthly ones. That is a lot of reach for a manageable amount of storage.

Finally the total is multiplied by the number of copies you keep away from the server, because every extra location holds the whole set again. The storage price, if you enter one, is applied to that total to give a monthly figure.

Complete copies or changes only

The two cards compare the main ways backup software stores data.

Complete copy every time. Each run saves everything. It is simple and robust. Any single copy can be restored on its own, and losing one does not damage the others. The price is storage and upload time: every run moves the whole site again, even if only a few files changed.

Weekly complete copy, changes in between. One complete copy is taken each week, and each run in between saves only what changed since the previous run. This is often called an incremental backup. It uses far less space and finishes much faster, which is what makes frequent backups practical for a larger site.

The trade-off shows up when you restore. To get to Friday's state you need the last complete copy, then Tuesday's changes, then Wednesday's, then Thursday's, then Friday's, applied in the right order. If one of those pieces is missing or damaged, everything after it is unreachable. A restore also takes longer, because more has to be downloaded and applied. The tool counts the pieces and adds up the data for the worst case, which is the day before the next complete copy is due.

Reading the results

One complete copy. The compressed size of your site. Every other figure is built on it.

Most you could lose. The gap between backups. It assumes everything works as planned.

If one backup fails. Backups do fail, often quietly: a full disk, an expired password, a job that stopped running months ago. If the most recent one is missing, you lose twice the interval. This figure is a reminder that the schedule on paper is the best case, not the guarantee.

Storage saved by changes-only. How much smaller the retention set is when only changes are stored between weekly complete copies. If the saving is small, the change rate is high enough that the simpler approach costs little extra.

Each scheduled run. The size of an ordinary backup and the time to upload it at your speed. This turns red when the upload takes longer than the gap before the next run, which means jobs would overlap, queue up or be skipped.

Worst-case restore. Download time for everything needed, plus the overhead you set. This is a floor, not a promise.

Why upload time is the quiet problem

A site that backs up comfortably at launch can outgrow its schedule without anyone noticing. A nightly job that moved two gigabytes in minutes now moves forty and runs into the morning, then into the next job.

Speeds are quoted in megabits per second, while sizes are in gigabytes. Eight megabits make one megabyte, so a 50 Mbps upload moves at most about 22 gigabytes an hour in theory. In practice a backup job competes with the website itself, other processes and the provider's own limits. The "usable share of the link" field accounts for that. Seventy percent is a reasonable starting point. If your backups run while the site is busy, lower it.

The restore overhead nobody counts

Transfer time is only part of a restore. The overhead field is for everything else: finding the right copy, creating a clean environment, importing the database, repointing the site, clearing caches, and checking that pages, logins, forms and checkouts behave.

Forty-five minutes is a modest starting figure for a small, well-understood site. It is optimistic for anything with a shop, a membership area or custom code, and optimistic for anyone doing a restore for the first time under stress. A database import alone can take longer than the download, because the server has to rebuild every table and index.

The 3-2-1 idea and the copies setting

The copies field reflects a long-standing rule of thumb: keep at least three copies of anything you cannot afford to lose, on two different kinds of storage, with one stored somewhere else entirely. For a website that usually means the live site, a copy on or near the server for quick restores, and a copy at an independent location for disasters.

The rule exists because backups kept in the same place as the site share its fate. A hacked account, a billing suspension, a failed disk array or a deleted hosting plan can take the site and its on-server backups together. An independent copy survives all of those.

Files and database are different jobs

A site's files and its database usually have different change patterns, and the planner treats them as one figure for simplicity. In reality the database is often the smaller part and the one that changes constantly: every order, comment and login writes to it. Uploaded media is the opposite: large, and changing only when someone adds a file.

That difference is worth using. Some people back up the database every hour and the media folder once a day. If that is your setup, run the tool twice, once for each part, and add the results. The figures will be more honest than a single blended percentage.

A database should be backed up in a way that produces a consistent snapshot, normally through an export or a tool designed for the purpose. Copying the raw database files while the server is writing to them can produce a backup that looks fine and fails to load. Our article on what a website backup actually saves covers this and what else a complete backup should include.

Choosing a schedule

There is no single right answer, but the pattern is consistent.

A site that rarely changes, such as a brochure site or a portfolio, is well served by a daily or weekly backup. The most you could lose is some edits you can redo.

A site that takes orders, bookings or paid memberships carries data that cannot be recreated. Losing six hours of orders means customers who paid and are not in your records. For these, backups every few hours, or continuous database logging if your platform supports it, are worth the extra storage. Pick the interval from how much you would lose, not from what the provider's default happens to be.

A busy community site produces a stream of small changes from many people. Frequent changes-only backups for the database, with the media backed up daily, usually give the best balance.

Whichever you pick, check the result on the red-flag lines. If the run time is longer than the gap, or the restore estimate is longer than the time your site can be down, the schedule is not a plan yet.

How to test the plan you chose

A figure on a screen is not evidence that a restore works. Three tests turn the estimate into something you can rely on.

  1. Restore to a scratch location. Use a temporary subdomain or a local copy, never the live site. Time the whole process from the moment you start looking for the backup.
  2. Check what comes back. Open pages, log in, submit a form, run a test order, look at recent uploads. A restore that loads the home page proves little.
  3. Replace the estimate. Enter your real restore time as the overhead and see whether the result still meets the downtime you can accept.

Repeat after major changes: a new shop, a migration, a switch of hosting or backup tool. Backups tend to break at exactly those moments.

If you are still deciding where a site should live, the hosting cost calculator shows how plan type changes the bill, and the bandwidth estimator helps you see whether offsite transfer fits within your limits. Many plans count outbound data, including backups sent elsewhere.

Backups and the rest of the renewal calendar

Backup storage is a recurring cost like any other, and its price can change at renewal. If you pay for a backup service or a storage plan, add it to the renewal cost calculator with the rest of your domain, hosting and SSL, so you see it in the same yearly total and get a reminder before it renews.

Uptime and backups answer different questions. An uptime guarantee says how much downtime a provider tolerates before it owes you a credit. A backup says whether you can get your data back at all. The uptime SLA calculator puts numbers on the first. This tool does it for the second.

What this tool does not know

It only knows the numbers you give it. It cannot see your server, measure your real change rate, test your connection or confirm that a backup is valid. It assumes changes spread evenly across the day and that a day's changes are all new data, so a site that rewrites the same records repeatedly will be smaller in practice than shown. It does not include encryption overhead, provider request charges, egress fees for downloading your own data, tax or currency conversion.

It does not recommend a backup product or a storage provider, and none is named in it.

Use it when

  • Choosing how often to back up and how long to keep copies before you buy a backup service.
  • Finding out why a backup storage bill is growing, and how much retention you actually need.
  • Checking that a nightly or hourly job can finish before the next one starts.
  • Estimating how long a full restore would keep your site offline.
  • Comparing complete copies with changes-only backups for your own site size and change rate.

Do not rely on it for

  • Treating the estimate as proof that your backups work. Only a real test restore shows that.
  • Sizing backups for a database-driven shop from a single blended change rate. Plan the database separately.
  • Counting copies stored in the same hosting account as independent off-site copies.
  • Using the restore time as a guarantee. Provider waits, large databases and human error all add to it.
  • Replacing the retention and recovery rules your industry or your contracts require.

Common questions

How often should I back up my website?

Back up as often as you can afford to lose the data in between. A site that rarely changes is fine with a daily or weekly backup. A shop, booking site or membership site that records payments or customer data needs one every few hours or better. The planner's 'Most you could lose' figure is that gap.

What is the difference between a complete and an incremental backup?

A complete backup saves everything every time and can be restored on its own. An incremental backup saves only what changed since the previous run, so it is smaller and quicker, but a restore needs the last complete copy plus every later increment in order.

How much backup storage do I need?

Take the compressed size of your site and multiply it by the number of backups you keep, then by the number of off-server copies. Changes-only backups reduce that a lot when your site changes slowly. Enter your numbers and the planner works it out for both approaches.

What is the 3-2-1 backup rule?

Keep three copies of your data, on two different kinds of storage, with one copy kept somewhere else entirely. For a website that usually means the live site, a copy near the server for quick restores, and an independent off-site copy for disasters.

Do I need to back up the database separately from the files?

Often yes. The database changes constantly and is usually small, while uploaded media is large and changes rarely. Backing them up on different schedules is common, and it should be done with an export that produces a consistent snapshot.

How long should I keep old backups?

Long enough to notice a problem that started quietly. A week or so of frequent backups, a month of weekly copies and a few monthly copies cover most small sites. Beyond a year, the extra storage rarely helps unless a contract or regulation requires it.

Is the information I enter stored anywhere?

No. The calculation runs in your browser and nothing is sent to a server or saved. Reloading the page resets the figures to the defaults.

Does this tool recommend a backup provider?

No. It names no provider and takes no payment from any. The example figures are generic and are there to show how the numbers behave.

Other tools