Backups & Data Recovery
Every SpliceRun plan includes daily backups, self-service restore, and clear visibility into when the last backup actually ran. This page explains how it works day-to-day.
Daily backups, on every plan
- Your instance is backed up automatically once a day - no setup, no add-on required, on every plan including Bring your own server.
- Your dashboard always shows exactly when the last backup ran, not a vague "we back up regularly" claim. If a scheduled backup fails, you'll see it on your dashboard immediately - it never passes silently.
- You can restore from any available backup yourself, from your dashboard, without waiting on support.
What a backup actually contains
A usable backup is three things, and one missing piece makes the whole thing worthless. Every snapshot captures all three:
| Why it matters | |
|---|---|
| The database | Your workflows, credentials and execution history |
| The data volume | Configuration, custom nodes and binary data |
| The encryption key | Without it, every stored credential restores as unreadable text |
That third item is the one most commonly missed by hand-rolled backups. A database dump on its own will restore your workflows with every credential broken.
How they're stored
Backups are encrypted before they leave the server and only ever stored encrypted. They're deduplicated, so keeping many restore points costs far less space than keeping many full copies.
On Bring your own server, backups are pulled from your machine over the same secure connection we use to manage it. Nothing is written to your disk in the process.
Restoring your instance
Restoring puts your instance back to the state captured in the backup you choose - same workflows, same credentials, same execution history.
Before anything changes, we take a fresh snapshot of your current state. That's your undo. Then:
- Your instance is stopped
- The chosen snapshot is applied - database, volume and encryption key together
- Your instance is started again
- We health-check it
If it doesn't come back up healthy, we automatically put it back exactly as it was and tell you the restore was rolled back. A restore going wrong should never be the thing that takes you offline.
Back up before a risky change
If you're about to make a change you're unsure of, take a manual backup first. The button is on your instance page, and the snapshot is available to restore immediately.
This is the single most useful habit for anyone building automations, and it costs nothing.
Retention
Every plan is backed up daily, and every plan keeps the same older restore points: 4 weekly and 3 monthly. What differs is how many daily restore points you keep:
| Plan | Daily restore points | Weekly | Monthly |
|---|---|---|---|
| Shared (XS and S) | 7 days | 4 | 3 |
| Bring your own server | 7 days | 4 | 3 |
| Dedicated | 30 days | 4 | 3 |
So on Shared you can roll back to any of the last 7 days, then to any of the last 4 weeks, then to any of the last 3 months. On Dedicated the daily window is 30 days instead of 7.
Longer daily retention is available on any plan as an add-on - 30 or 90 days - see the pricing page. If you buy one, it replaces the plan's daily window rather than adding to it, and the longer of the two always wins.
Why Shared is backed up at all
Shared does not mean shared data. Your instance has its own database and its own storage volume, not a slice of a communal one, and a backup dumps exactly those - never a snapshot of the whole machine. That is what makes a per-customer restore possible on Shared, and it is why we include backups on the $5 plan rather than treating them as a premium feature.
Every snapshot is encrypted before it leaves the server, and includes your instance's encryption key, so a restore brings your stored credentials back working rather than unreadable.
Exporting your data
Any restore point can also be exported - a single archive holding the database, the data volume and the encryption key, downloadable for 24 hours. That is everything needed to run your instance somewhere else, and you can take one whenever you like without affecting your running instance or using up a restore point.
See Exporting Your Data for what the archive contains and how to import it, and Billing, Cancellation, and Data Policy for how exports interact with cancellation timelines.
Higher availability
Daily restore points suit most workloads, and they are what we offer today.
We do not currently sell a hot standby that takes over automatically. It was listed as an add-on and has been withdrawn, because the failover it implied was not built - we would rather not sell it than sell it and hope. If your workload genuinely cannot tolerate the time a restore takes, tell us before you buy and we will be straight with you about whether SpliceRun is the right host for it today.
What you can do now: take a manual backup before risky changes, keep a longer daily retention window as an add-on, and export your data whenever you like.
What backups can't do
They capture a point in time. Data created and then deleted between two snapshots was never in a backup, and can't be recovered.
They're not a live replica. See above if you need continuous protection.
On Bring your own server, they don't protect the server itself. They protect the application and its data. If the machine is lost you can restore onto a new one - but provisioning that new machine is yours to do.
If SpliceRun ever discontinues a service
Every active customer receives a full backup of their instance plus the standard open-source runtime for whichever app it runs, so you can self-host it elsewhere with reasonable advance notice. Everything we host is open-source software, so this costs us nothing to promise - it's a real anti-lock-in commitment, not a marketing line.
Questions about backups or a specific restore? Contact Support.