Overview
What the Scaleway integration does
The Scaleway integration backs up Scaleway Instances, Block Storage volumes and
Secret Manager secrets into an encrypted, deduplicated Plakar store, without
installing anything inside the Instances. It restores them as new resources in
the same project, another project or another zone. It runs from the free
plakar CLI or from Plakar Control Plane, which also discovers the resources of
each Scaleway project. Everything goes through the Scaleway APIs and the Object
Storage endpoint over HTTPS, so Plakar runs inside Scaleway or on any other
infrastructure, at another cloud provider or on premises.
Three protocols cover the project. scaleway-instance:// backs up an Instance
with every Block Storage volume attached to it. scaleway-block:// backs up a
single volume. scaleway-secret:// backs up every secret of a project and
region, with all their versions. Object Storage buckets and Managed Databases
for PostgreSQL and MySQL go into the same store through the
S3, PostgreSQL and
MySQL integrations.
$ plakar pkg add scaleway
$ plakar source add api-prod \
scaleway-instance://5b0f3c3e-7a51-4d6e-9c47-1f2b8d0e6a91 \
access_key=SCWEXAMPLE7KQ2 secret_key=$SCW_SECRET_KEY \
project_id=$SCW_PROJECT_ID bucket=plakar-staging zone=fr-par-1
$ plakar at /var/backups backup "@api-prod"
c41e9b07: OK ✓ /.METADATA.json
c41e9b07: OK ✓ /block-storage/0c9d4e1a.api-prod-root/.METADATA.json
c41e9b07: OK ✓ /block-storage/0c9d4e1a.api-prod-root/0c9d4e1a.qcow2
c41e9b07: OK ✓ /block-storage/7e21b5f8.api-prod-data/.METADATA.json
c41e9b07: OK ✓ /block-storage/7e21b5f8.api-prod-data/7e21b5f8.qcow2
c41e9b07: OK ✓ /
c41e9b07: backup completed without errors in 6m48.112s
$ plakar at /var/backups ls
2026-11-13T01:00:03Z 7f3a20d1 40 GiB 6m31s /
2026-11-14T01:00:04Z c41e9b07 40 GiB 6m48s /
$
Agentless
Why use Plakar for Scaleway backup
Nothing to install in your Scaleway Instances
Plakar works through the Scaleway APIs only: it snapshots the volumes, exports them as QCOW2 images and reads them from Object Storage. No package, no cron job and no open port inside the Instance, and the same approach covers Linux and Windows Instances alike.
Instances, volumes and secrets in one store
One integration protects the compute, the data disks and the secrets the
applications read at startup. They land in the same encrypted store, with the
same retention and the same plakar check.
Restores that add, never overwrite
A volume comes back as a new volume, an Instance as a new Instance or as extra disks on an existing one, a secret as new revisions. The resource you restore next to stays exactly as it is.
Another zone or another project
Each destination sets its own zone and project_id, so the same snapshot
restores in the original project, a recovery project, or another Scaleway zone.
Recovery scenarios
Scaleway backup and recovery scenarios
An infrastructure agent deletes a production Scaleway Instance
Asked to remove unused development servers, an AI agent with the operator’s API
key deletes api-prod-1 and its volumes. The volume snapshots of the project
are managed with that same key. Run plakar backup before handing the project
to the agent, and keep the store and its passphrase out of the agent’s
environment. Restore with a scaleway-instance://spawn destination: Plakar
recreates the Instance with its name, commercial type, tags and boot layout, and
starts it.
A Scaleway API key leaks from a CI job
A full-access key for the production project is printed in a build log and used
to delete Instances and their snapshots. Keep the Plakar store in another
project with its own key, or at another provider, verify it with plakar check,
and restore the Instances into a clean project by setting project_id on the
destination.
Recovering Scaleway Instances in another zone
When an Availability Zone is unavailable, restore the latest snapshot as a new Instance in another zone or region. Plakar exports and imports the QCOW2 images for you, so the move takes one restore command. Attach a flexible IP or a Private Network to the new Instance and switch traffic to it.
A Scaleway secret deleted two weeks ago
A cleanup script deleted the secret holding the production database password. Secret Manager keeps a deleted secret recoverable for 7 days, then removes it with all its versions. A Plakar snapshot keeps every version for as long as your retention policy says, and the restore adds them back as revisions, enabled or disabled as they were.
One overwritten file on a Scaleway data volume
A deployment replaced /etc/app/config.yaml on the data volume yesterday.
Restore last night’s volume snapshot as a new volume attached to the same
Instance, mount it read-only, copy the file back and detach the volume. The live
volume is never touched.
How it works
How Plakar backs up a Scaleway Instance without an agent
Scaleway exposes the content of a volume by exporting a snapshot of it to Object Storage as a QCOW2 image. Plakar drives that export through the API and streams the image into the store. A dedicated bucket of the project serves as the staging area, and Plakar removes what it created there.
Each snapshot holds the Instance description returned by the Scaleway API and,
for every Block Storage volume, its description and its QCOW2 image. Browse it
with plakar ls, read a description with plakar cat, or plakar mount the
snapshot to open the images with standard QEMU tools.
Snapshot. Plakar creates a Block Storage snapshot of each volume attached to the Instance.
Export. Scaleway exports each snapshot as a QCOW2 image into the staging bucket, and Plakar deletes the temporary snapshot.
Store. Plakar reads each image, encrypts and deduplicates it into the store, then deletes the image from the bucket.
c41e9b07:/
├── .METADATA.json # the Instance: name, type, tags, zone
└── block-storage/
├── 0c9d4e1a.api-prod-root/
│ ├── .METADATA.json # the volume: name, size, IOPS, tags
│ └── 0c9d4e1a.qcow2
└── 7e21b5f8.api-prod-data/
├── .METADATA.json
└── 7e21b5f8.qcow2Complementary
Scaleway snapshots and Plakar: what each one covers
Scaleway’s own features and Plakar work side by side. Keep the native ones for fast rollback inside a project, and add Plakar for a history you hold elsewhere.
Block Storage snapshots for a quick rollback in the zone.
Plakar adds a history of exported QCOW2 images, deduplicated from one night to the next, in a store that does not depend on the project’s snapshots.
Snapshot export and import to move a volume between regions.
Plakar runs the export on every backup and the import on every restore, so a
move to another region is a destination with a different zone.
Object Storage with Object Lock for WORM retention.
Plakar writes its store to a bucket like any S3 client, and Object Lock keeps every object for the retention set on the bucket. The 3-2-1 layout below uses it for the archive copy.
Secret Manager versions and a 7-day recovery window.
Plakar keeps every version of every secret for the retention you set, and restores them into any project and region.
Managed Database automated backups.
Plakar connects to the database engine with a database user, from any host that reaches it, with no staging bucket, and restores into the same Managed Database, another one or a self-managed server.
3-2-1
A 3-2-1 Scaleway backup with Object Lock
Three copies, on two storage systems, one of them in another region and locked.
| Copy | Where | Role |
|---|---|---|
| 1. Production | Instances, volumes, buckets, Managed Databases and secrets in fr-par |
The data the applications use |
| 2. Primary store | An Object Storage bucket in fr-par |
Day-to-day restores |
| 3. Archive store | An Object Storage bucket in nl-ams with Object Lock in compliance mode, under its own IAM application or project |
The off-site copy, fed by plakar sync |
$ plakar store add scw-prod location=s3://s3.fr-par.scw.cloud/plakar-prod \
access_key=$SCW_PROD_ACCESS_KEY secret_access_key=$SCW_PROD_SECRET_KEY
$ plakar store add scw-archive location=s3://s3.nl-ams.scw.cloud/plakar-archive \
access_key=$SCW_ARCHIVE_ACCESS_KEY secret_access_key=$SCW_ARCHIVE_SECRET_KEY
$ plakar at "@scw-prod" create
$ plakar at "@scw-archive" create
$ plakar at "@scw-prod" backup "@api-prod"
$ plakar at "@scw-prod" sync to "@scw-archive"
$ plakar at "@scw-archive" checkEach synced snapshot restores on its own from nl-ams, so losing fr-par
leaves every restore point of the archive available.
Comparison
Scaleway backup tools compared
| Plakar | Commercial backup suites | Open-source backup tools in the Instance | Scripts with the Scaleway CLI | |
|---|---|---|---|---|
| Nothing installed in the Instances | YesScaleway APIs only | NoAn agent in each Instance | NoInstalled in each Instance | YesThrough the API |
| Copies kept outside the Scaleway project | YesExported to a store you choose | YesIn the product’s repository | YesIn the bucket you point them to | PartlyIf your script exports them |
| Instances with all their volumes | YesOne scaleway-instance:// source |
PartlyThe files the agent sees | PartlyThe directories you list | PartlyYour script |
| Secret Manager secrets and versions | Yesscaleway-secret:// |
NoNo Secret Manager support | NoNo Secret Manager support | NoNo ready-made script |
| Encrypted, deduplicated history | YesVerified with plakar check |
PartlyEncryption as an option | YesEncrypted, incremental | NoYour script |
| Restore into another zone or project | YesSet zone and project_id |
PartlyThrough the agent | PartlyFiles restored into an Instance you create first | PartlyExport, copy and import by hand |
| Managed Databases, buckets and Kapsule clusters in the same store | YesPostgreSQL, MySQL, S3 and Kubernetes integrations | NoHosts with an agent only | NoFiles of the Instance only | PartlyYour script |
| Discovery of the project’s resources | YesScaleway inventory in Plakar Control Plane | NoHosts added one by one | NoSet up on each Instance | NoYour own tooling |
| Free CLI | Yes | NoCommercial licence | Yes | Yes |
Restore
How to restore a Scaleway Instance in another zone or project
A restore never touches the backed-up Instance. Pick a destination, point it at the target zone and project, and restore any snapshot into it.
Find the snapshot. List the snapshots of the Instance with plakar ls and
pick the last good one.
Prepare the target. Create a staging bucket in the target region and an API key with the same permissions in the target project.
Restore. Add a scaleway-instance://spawn destination with the target
zone and project_id, then run plakar restore.
Reconnect. Attach a flexible IP or a Private Network to the new Instance.
$ plakar at /var/backups ls
$ plakar destination add api-prod-ams scaleway-instance://spawn \
access_key=SCWEXAMPLE7KQ2 secret_key=$SCW_SECRET_KEY \
project_id=$SCW_RECOVERY_PROJECT_ID \
bucket=plakar-staging-ams zone=nl-ams-1
$ plakar at /var/backups restore -to "@api-prod-ams" c41e9b07To add the disks to an Instance that already exists, use
scaleway-instance://<server-id> as the destination: each volume of the
snapshot is attached to it as a new volume.
Secrets
Scaleway Secret Manager backup
Every secret, every version
A backup reads all the secrets of a project and region, with each version and its enabled or disabled status.
Restore into any project and region
The destination sets its own project and region. A secret with the same name and path is reused, a missing one is created, and every version is added as a new revision.
Scaleway Secret Manager as a secret provider
Plakar Control Plane can read the credentials it uses from Scaleway Secret
Manager at run time, for example
c492b67c-f3a0-45b8-9ae5-87b640354fe5#access_key.
Quickstart
How to back up Scaleway with the open-source Plakar CLI
The Plakar CLI is free and open source. The Scaleway integration is a pre-built package, free to install once you log in with a Plakar account. Create a store, declare a source per Instance, volume or project of secrets, and run backups from cron or a CI job.
The Scaleway CLI documentation lists every option, the permissions of the API key for each protocol and the volume restore.
$ plakar login -email you@example.com
$ plakar pkg add scaleway
$ plakar at /var/backups create
$ plakar source add api-prod \
scaleway-instance://5b0f3c3e-7a51-4d6e-9c47-1f2b8d0e6a91 \
access_key=SCWEXAMPLE7KQ2 secret_key=$SCW_SECRET_KEY \
project_id=$SCW_PROJECT_ID bucket=plakar-staging zone=fr-par-1
$ plakar at /var/backups backup "@api-prod"
$ plakar source add prod-secrets scaleway-secret:// \
access_key=SCWEXAMPLE7KQ2 secret_key=$SCW_SECRET_KEY \
project_id=$SCW_PROJECT_ID region=fr-par
$ plakar at /var/backups backup "@prod-secrets"
$ plakar at /var/backups checkControl Plane
Scaleway backups across projects with Plakar Control Plane
With several projects and teams, Plakar Control Plane replaces per-host CLI configuration. One Control Plane protects the Instances and volumes of several projects through the Scaleway API, with no network peering. A Scaleway inventory per project discovers its Instances, Instance volumes, Block Storage volumes, Object Storage buckets, managed databases and Secret Manager, and keeps the list current as resources come and go. Restores run from the same interface, as new volumes, new Instances or new secret revisions.
See how Plakar Control Plane works, read the Control Plane Scaleway documentation and the installation on Scaleway, or contact us to plan a deployment.
Connect a project. Create a Scaleway inventory with a read-only API key
scoped to the project (AllProductsReadOnly). Keep separate IAM applications
for discovery, backup and secrets, so the full-access keys only serve the jobs
that need them.
Keep credentials in Secret Manager. Reference the backup key as
{secret_id}#{field} so Plakar Control Plane resolves it at run time.
Let a policy schedule it. Policies set how many backups each resource needs per period, how long they are kept and which store receives them. Every Instance, volume or secret project that matches is scheduled automatically, including the ones created next week.
FAQ
Scaleway backup FAQ
What does Plakar install inside Scaleway Instances?
Nothing. Plakar works through the Scaleway APIs: it snapshots the Block Storage volumes, exports them to Object Storage as QCOW2 images and reads them from there.
How does Plakar back up a Scaleway Instance?
It snapshots every Block Storage volume of the Instance, exports each snapshot to a staging bucket as a QCOW2 image, and stores the images with the Instance and volume descriptions in an encrypted, deduplicated snapshot.
What permissions does the Scaleway API key need?
For Instances and volumes, full access to Instances and Block Storage, and read and write access to Object Storage buckets and objects. For secrets, Secret Manager read, access, create, write and restore permissions.
Can I restore a Scaleway Instance in another zone or project?
Yes. The destination sets its own zone and project_id, and
scaleway-instance://spawn creates the Instance there from the snapshot.
How do I back up a single Scaleway Block Storage volume?
Declare it as a scaleway-block://<volume-id> source. A restore brings it back
as a new volume, detached with scaleway-block:// or attached to an Instance
with scaleway-block://<instance-id>, with the name, tags and IOPS of the
original.
What does a Scaleway restore create?
New resources only: a new volume, a new Instance, or new revisions of a secret. The resource you restore next to is left as it is.
What happens to the staging bucket?
Plakar deletes each QCOW2 image once it has been read, and each temporary Block Storage snapshot once it has been exported.
How long does Scaleway keep a deleted secret?
Secret Manager keeps a deleted secret recoverable for 7 days, then deletes it with all its versions. Plakar keeps backed-up versions for as long as your retention policy sets.
Can Scaleway secrets be restored into another project?
Yes. The destination sets its own project_id and region. Secrets with the
same name and path are reused, and every version is added as a new revision.
Where should I store Scaleway backups?
In a store outside the project you protect: an Object Storage bucket in another
project or region, with Object Lock enabled, or a store at another provider.
plakar sync keeps a second copy.
How do I back up Scaleway Managed Databases for PostgreSQL and MySQL?
With the PostgreSQL and MySQL integrations, which connect to the managed server with a database user and store the dumps in the same Plakar store. A restore goes to the same Managed Database, another one or a self-managed server.
How do I back up Scaleway Object Storage buckets?
With the S3 integration, which reads the bucket over the Scaleway S3 endpoint,
for example s3://s3.fr-par.scw.cloud/my-bucket. A bucket can also host the
Plakar store or receive a restore.
How do I back up Scaleway Kubernetes Kapsule clusters?
With the Kubernetes integration, which connects with the cluster’s kubeconfig, backs up every resource as YAML, namespace by namespace, and reads the data of PersistentVolumeClaims through a helper pod.
Where can Plakar run to back up Scaleway?
Anywhere that reaches the Scaleway APIs and the Object Storage endpoint over HTTPS: an Instance of the protected project, another Scaleway project, another cloud provider or an on-premises server. The disks transit through a staging bucket of the protected project wherever Plakar runs.
How is the Scaleway integration distributed?
As a pre-built package, free to install with a Plakar account, with
plakar pkg add scaleway.
Should I use the free CLI or Plakar Control Plane?
Start with the free CLI for a few Instances run from cron. Move to Plakar Control Plane for discovery across projects, policies and one view of what is protected.