Refer to the Codelax repository for a demo on how to use the plugin.

Refer to the Scaleway Ansible module repository for more information.

Scaleway is now listed on the official Upbound repository

Refer to the Scaleway Crossplaine Provider repository for more information.

HTTP redirects come as a new type of ACL configured within a frontend. You can fully customize the redirection URL, from the protocol scheme all the way to the query parameters, using powerful templated placeholders to reuse parts of the original requests.

Furthermores, we are introducing the ability to create unconditioned ACLs, allowing you to explicitly define a default and catch all behavior for your frontend’s ACL.

Check out our documentation to find out how to create and manage your ACLs to get started, or our reference documentation for a deeper dive into ACLs.

Refer to our Mimir metrics documentation to find out about the limitations for Mimir.

If you need custom limitations, feel free to reach out to us.

To deploy an example application with the self-hosted API gateway, refer to our official GitHub examples repository.

Check out our documentation to start setting up granular access to Object Storage.

• Node20 in beta. It includes the latest version of nodeJs as well as all the improvements that were added to node19.

• TransactionalEmailDomainFullAccess
• TransactionalEmailDomainReadyOnly
• TransactionalEmailEmailFullAccess
• TransactionalEmailEmailReadyOnly

For more information, refer to the Secret Manager product documentation or the Secret Manager product website

How it works: The monthly pricing for custom metrics will be calculated in levels. Scaleway applies six levels of volume discounts.

Here is an example of how you would be billed for sending 52 billion custom metrics samples per month:

• 10 billion samples at €0.15 per million samples: €1.500
• 15 billion samples at €0.13 per million samples: €1.950
• 25 billion samples at €0.12 per million samples: €3.000
• 2 billion samples at €0.11 per million samples: €220

Total: €6.650 per month with volume discounts instead of €7.800 without volume discounts.

Upgrade your monitoring experience with Cockpit and enjoy smarter pricing that scales with your needs!

Gain insights into every aspect of your email delivery with a comprehensive array of metrics, including:

• Current domain reputation: Stay informed about your domain's reputation score.
• Reputation score evolution: Track how your reputation score evolves, and make informed decisions to maintain or enhance it. Read our dedicated documentation to find out how.
• Total emails sent: Keep track of the volume of emails your system is processing.
• Total email drops: Identify delivery issues and take proactive measures to address them.
• General activity overview: Get a bird's-eye view of your email activity trends over time.
• Error activity: Pinpoint any errors or issues in your email delivery pipeline to solve them swiftly.
• Top recipient destinations: Understand who your emails reach the most.
• Top bounce and fail recipients: Identify problematic recipients and take action to rectify issues.

With these metrics at your disposal, optimizing your email delivery and ensuring a seamless user experience has never been simpler.

• Traces come with a retention period of 7 days in the free pricing plan.
• Traces are billed 0.35€ per GB ingested, per month.

Get detailed records of your requests' behavior with traces, and enhance your monitoring capabilities!

Whether you are seeking reliability, speed, or localized service, you can now order your hosting from the location that best suits your needs.

How it works:

The migration process starts off from your Online Console. Simply initiate the process when prompted, and you will be guided through each step. If you have not already created a Scaleway account, you will receive an invitation to create one in the near future.

Our migration tool handles everything for you, ensuring a smooth transition. Plus, if you also have domains hosted with Scaleway/Online, the process becomes even more streamlined. All you need to do is select your new hosting offer, and let the tool take care of the rest.

Upgrade to Scaleway Elements today and experience the next level of Web Hosting!

If you want to know more about the changes behind this evolution, have a look at our blog post.

Find out more about the manual and automatic migrations plans.

Access your dashboard via your domain's Metrics tab in the Scaleway console.

• application/xml
• text/xml

Find out about the full range of available attachment types in our documentation.

Key features include:

• Nvidia L40S 48 GB (Ada Lovelace architecture)
• 4th generation Tensor cores
• 4th generation RT cores (graphics capability)
• Transformer Engine
• Available in multi-GPU 1, 2, 4 or 8 GPUs

• 50 SQS queues
• 50 SNS topics
• 50 NATS streams

This update empowers you to handle larger workloads and enables more robust communication.

For detailed information on the limitations and capabilities of our Messaging and Queuing protocols, please refer to the documentation.

• 50 SQS queues
• 50 SNS topics
• 50 NATS streams

This update empowers you to handle larger workloads and enables more robust communication.

For detailed information on the limitations and capabilities of our Messaging and Queuing protocols, please refer to the documentation.

• 50 SQS queues
• 50 SNS topics
• 50 NATS streams

This update empowers you to handle larger workloads and enables more robust communication.

For detailed information on the limitations and capabilities of our Messaging and Queuing protocols, please refer to the documentation.

The Mac mini M1 (M1-M) range will stay on macOS 13 by default for now. If you want to enjoy a smooth macOS 14 migration you can opt-in to our latest macOS configuration improvements by reinstalling your server through the console, then performing a manual upgrade using softwareupdate.

The initial commitment is for one month from the server order. At the end of your initial commitment period, the commitment is automatically renewed every month on the anniversary date unless you choose to delete the machine, in which case the deletion will take effect on the anniversary date of the commitment.

Monthly pricing:

• M1-M: €75/month
• M2-M: €115/month
• M2-L: €159/month
• M4-S: €149/month

Read our documentation on how to manage commitment plans.

The API v1 is now deprecated and will be definitively removed by October 1, 2025. Public Gateways users are encouraged to migrate to the API v2. Learn more in our detailed documentation.

Both models are multimodal and support text generation and image analysis use cases.

Check the full list of supported file extensions in the supported file types documentation.

This feature protects infrastructures and applications from the impacts of accidental deletion of secrets or malicious actions.

For more details, check out our API documentation for Product Catalog.

Key features include:

• Nvidia H100 SXM 80GB GB (Hopper architecture)
• 4th generation Tensor cores
• 4th generation NVlink, which offers 900 GB/s of GPU-to-GPU interconnect
• Transformer Engine
• Available immediately in 8 GPUs (2 and 4 GPUs per VM coming soon)

Bring Your Own Subscription (BYOS)

RHEL on Scaleway operates on a Bring Your Own Subscription model. You can deploy a RHEL Instance directly from the Scaleway console and activate it using your existing Red Hat subscription to receive updates, security patches, and support.

Prerequisites

• An active Red Hat subscription eligible for Red Hat Cloud Access.
• A Red Hat Customer Portal account.
• Standard Scaleway account permissions (Owner or IAM permissions).

Important Notes

• Licensing: Instances are deployed unentitled. You must register and attach your subscription to access Red Hat repositories.
• Support: Technical support is provided directly through your Red Hat subscription, not via Scaleway Support.
• Licenses: RHEL licenses must be ordered directly from Red Hat via the Customer Portal.

For all future feature requirements, we encourage you to open a specific feature request of what you need to integrate into Cockpit.

WAW-3 is now fully operational, providing enhanced accessibility and proximity to support our customers in the Central European region. This comes as the latest development in our commitment to offer an ever-resilient and high-performing cloud infrastructure. Deploy your Instances in WAW-3 now to experience unmatched performance, reliability, and data sovereignty.

For more information, refer to the How to manage Read Replica documentation page.

In addition, we have introduced a balanced resource allocation system, establishing a relationship between RAM and CPU to help you select optimal resource combinations.

Since quotas are applied at Organization level, this improvement makes it easier to monitor your resource consumption and anticipate quota requests.

For streaming responses, the usage field is incremented in each chunk, and completed in the very last chunk of the response.

More details and examples in our OpenAI API compatibility documentation.

Qwen3 235B is currently the best performing open model, and compares with state-of-the-art proprietary models for text and agentic use cases.

L4 GPUs are now available in a second Availability Zone in Paris (par-1, in addition to par-2), making it easier to build highly available inference infrastructure for your projects.

As a reminder, L4 GPU Instances are also available in the Warsaw region (waw-2).

Key features include:

• Nvidia L4 24 GB (Ada Lovelace architecture)
• 4th generation Tensor cores
• 4th generation RT cores (graphics capability)
• Available in multi-GPU 1, 2, 4 or 8 GPUs

With NVLink GPU-to-GPU communication, the 4-GPU option unlocks even greater possibilities and higher performance for your deployments. Now available in the Paris (PAR2) region.

Key features include:

• Nvidia H100 SXM 80GB GB (Hopper architecture)
• 4th generation Tensor cores
• 4th generation NVlink, which offers 900 GB/s of GPU-to-GPU interconnect
• Transformer Engine
• Available now in 2, 4 and 8 GPUs per VM (Additional stock deployments on-going)

For more details on how to monitor your jobs, consult the dedicated documentation.

GPT-OSS is a frontier open source model with significant reasoning capabilities and increased tokens throughput performance.

Following our model lifecycle policy, both models will remain accessible up to November 14th, 2025. After this date, requests will be routed automatically to their most recent version.

Learn how to use Private Networks with Serverless Containers

Learn how to use Private Networks with Serverless Functions

All tab in server creation: When ordering a server, the “All” tab displays a complete list of all available servers, regardless of their range or other filters.

Automatic AZ redirection: If the selected AZ is out of stock for a specific offer, the console will now redirect you to another AZ where it is available.

For running Instances, no network configuration changes are needed unless you modify the public IPv4 address.

If you change the public IPv4 address, the Instance will retain its existing static configuration and continue using the previous address. To enable DHCP for your Instance, follow our dedicated documentation before updating the public IPv4 address.

• service-dns-ip: IP used for the DNS Service (cannot be changed later)
• pod-cidr: Subnet used for the Pod CIDR (cannot be changed later)
• service-cidr: Subnet used for the Service CIDR (cannot be changed later)

• Change storage and node size independently to optimize your costs and adapt to your application's traffic.
• Scale up your storage volume from 5GB to 10TB with no downtime.
• Make Snapshots of your data in seconds for any size of the volume and recover a previous state of your data in case of accidental data alteration.

• fr-par-1
• fr-par-2
• nl-ams-1
• pl-waw-1

• fr-par-1
• fr-par-2
• nl-ams-1
• pl-waw-1

* or equivalent CPU

You will also be alerted when the bounce rate of your emails exceeds a certain threshold, allowing you to proactively manage your domain's reputation.

Access logs are invaluable tools that record detailed information about every interaction with your objects. These logs include essential data such as who accessed your objects, when they accessed them, and from which IP address. With access logs enabled, you can gain deeper insights into how your data is being used, and enhance your security and compliance efforts.

For more information, refer to Scaleway Documentation.

Dedicated control planes 4/8/16GB are available from the Scaleway console, the CLI, and Terraform (>= 2.28.0). Refer to our documentation for more information.

The query protocol remains a supported option and you have the flexibility to use the SDK's earlier versions or a custom implementation using the query transport protocol.

Scaleway's Billing API gives you an overview of your cloud consumption and allows you to monitor expenses.

What's New?

• Monthly Consumption History - access the complete history of your monthly consumption. Until now, only data from the current month was available.

• Granularity of Costs by Project, Product Category, and Month - we have introduced increased granularity, allowing you to view costs based on the project, product category, and month.

Refer to the Billing API documentation for more information.

You can find detailed specifications of the four configurations below:

Available in PAR-2 for launch and PAR-1 in the coming days

These limits are now more permissive to allow different use cases.

• NATS (NATS)
• Queues (formerly SQS)
• Topics & Events (formerly SNS)

• NATS (NATS)
• Queues (formerly SQS)
• Topics & Events (formerly SNS)

• NATS (NATS)
• Queues (formerly SQS)
• Topics & Events (formerly SNS)

You can now see estimated carbon (Co2) impact metrics for both Block Storage and Object Storage directly in the Scaleway console before ordering a product.

This integration allows you to make informed decisions to optimize the sustainability of your full-stack infrastructure.

Visible now in the Object Storage and Block Storage sections of the console and also in our Product Catalog API.

• Connect your favorite tools, like Airflow, directly to your Apache Spark™ cluster and submit jobs effortlessly.
• Connect your Spark™ clusters to any compatible data source.
• Install custom or classic libraries (such as SparkSQL) and explore your data visually with graphs in JupyterLab.
• Easy Machine Learning with TensorFlow and MLlib.
• Monitor and analyze cluster and job performance with the Apache Spark™ UI.
• Your Spark™ cluster is now securely accessible via Private Networks.

Block Storage SLA is on par with the standards set by industry leaders and supported by a seasoned monitoring practice, ensuring a seamless experience for our customers to run their business. Refer to the Block Storage Service Level Agreement (SLA) for additional details.

The new EM-A315X-SSD offers:

• CPU: Intel Xeon E5 1410 v2 (4C/8T, 2.8 GHz)
• RAM: 64GB
• Disks: 2 × 1TB SSD
• Network: 1 Gbps
• Availability Zones: Now in PAR-1, soon in PAR-2

• jpeg
• jpg
• png
• gif

Find out about all the attachment types available.

• Rust is now in General Availability
• PHP is now in General Availability
• Node19 is available in beta
• Go 1.20 is available in beta

• Public-only Kapsule users should transition to using the Private Network feature for enhanced security and performance.
• Kosmos users with Scaleway pools should stop using this annotation.

Please refer to the related CCM documentation.

Patch versions 1.29.1, 1.28.6, 1.27.10, 1.26.13, 1.25.16 have also been released.

Read our newest version support policy for more details on the 14-month support timeline.

To apply the patch, managed Kubernetes services customers should take the following actions:

• Kubernetes clusters running on 1.24: replace nodes
• Kubernetes clusters running on 1.25+: upgrade to the latest k8s patch release

This feature is tailored for you! Enable your dedicated IP during your hosting order or later via the console, and we will provide you with a dedicated IP linked directly to your hosting plan.

• Replace or reboot your nodes.
• Use a fully qualified domain name (FQDN) for resolving addresses within the Private Network. For example, if you have a service named foo in the Private Network mypn, use foo.mypn.

Object Storage SLA is on par with the standards set by industry leaders, and supported by a seasoned monitoring practice, ensuring a seamless experience for our customers to run their business. Refer to the Object Storage Service Level Agreement (“SLA”) to know more.

• Improved resilience and scalability
• No more folder endpoints, but still the possibility to use paths to organize secrets
• Changing the path of a secret has become simpler

The Scaleway console now uses this version for the Secret Manager product pages. For more information, you can refer to the API documentation.

Note that the v1alpha1 API is still supported but it will eventually be deprecated.

Our brand-new Billing Space is now live and ready for you to explore. Accessible right from the console's right-hand menu, this space is designed to streamline your billing experience. Here is what is in for you:

• Overview Page: Your go-to destination for essential information on managing your usage and invoices.
• Consumption Management: Track your budget alerts and soon, the Cost Manager tool for cost tracking.
• Optimization Page: Access vouchers and upcoming cost optimization offers.
• Billing and Payments Information: A dedicated section for all your billing and payment details.

Explore these new features today and manage your billing more efficiently than ever before!

<Message type="tip">
  Refer to the [Billing documentation](/billing/) for more information.
</Message>

Key Manager helps organizations achieve secure key management by handling low-level and error-prone cryptographic details for you.

Find out more about Key Manager in the dedicated documentation.

DeepSeek R1 is an open-weight reasoning model matching proprietary models performance. Distilled version improves Llama model performance on reasoning tasks like mathematics or code.

DeepSeek R1 Distilled Llama improves Llama model performance on reasoning use cases like mathematics or code.

• TransactionalEmailEmailSmtpCreate: Grants the ability to create emails via SMTP.
• TransactionalEmailEmailApiCreate: Grants the ability to create emails via the API.

View the v1beta1 documentation on the Edge Services API reference page.

Learn more in our dedicated documentation: Monitor Data Plane with Cockpit

Additionally, you now get a breakdown of pricing changes when modifying resources.

Your container can now scale up or down based on CPU or RAM usage to better adapt to your specific workloads.

You can now track the environmental impact of your Apple servers with the same level of transparency as Elastic Metal.

The estimated impact is now displayed directly on the order page. Monthly environmental reports, including CO₂ emissions and water usage, are also available.

Start tracking your impact today! Check out your environmental reports in the Scaleway console.

The issue has now been fixed.

New pricing and specifications

These offers are available immediately. Login to your dashboard to deploy a new VPS today!

The impacted endpoints are logcenter.online.net and stats.online.net.

If you are currently using these services, we invite you to migrate your hosting to Scaleway to continue benefiting from logging and statistics features in our updated environment.

• The command field has been upgraded to startup_command and args for improved flexibility and more use cases
• The reason field was added to give more context about how a job is terminated

For more details about the changes, read our dedicated documentation.

Please, update your dependency: 

npm i serverless-scaleway-functions@~0.5.0

The redeploy: false option in the UpdateContainer API (PATCH /containers/v1beta1/regions/{region}/containers/{container_id}) is now deprecated. This parameter was previously used to prevent redeployment of containers when no configuration changes occurred. Going forward, the API will systematically redeploy containers when necessary, and any use of redeploy: false will be ignored.

We strongly advise against relying on this field in your automation or deployment workflows. If you need to force a redeployment regardless of configuration changes, please use the dedicated DeployContainer method instead.

IAM must be used to generate an API key with the permissions to manage Projects.

Find out more on the Managing projects using Scaleway APIs documentation page.

You can now create and manage policies to give fine-grained access rights to Scaleway users, applications and groups. API keys can be attached to a user or an application and inherit their rights.

IAM features are available though Console, CLI and Terraform.

For Organizations created before December 5th, IAM can be activated by the Owner of the Organization on the console. Accounts created since December 5th 2022 have IAM activated by default.

Find out more in our dedicated IAM documentation.

• Ubuntu 22.04 Jammy Jellyfish
• GitLab InstantApp
• Nextcloud InstantApp

• Centos Stream 9
• Fedora 36
• AlmaLinux 8
• Rocky Linux 8
• Debian Buster & Debian Bullseye

Users wanting to get the patch without having to launch new Instances can run the following commands:

• For Ubuntu and Debian: apt-get install scaleway-ecosystem
• For CentOS, Fedora...: dnf update scaleway-ecosystem

• Absence of configuration of the Private Network interface
• Misconfiguration of the Private Network interface
• The Private Network does not come up when hot-plugged on Ubuntu Focal

PRO2 Instances are now available in:

• fr-par-1
• fr-par-2
• nl-ams-2
• pl-waw-1

• Ubuntu 16.04 Xenial Xerus
• Centos 7.6
• CentOS 8
• Fedora 32
• OwnCloud Instant App

For API users, note that the local volume min_size returned by a call to /instance/v1/zones/{zone}/products/servers is now 0, whereas it was previously equal to the max_size.

Be careful if you are using this value in your automation scripts, as trying to create an Instance with a local volume of size 0 will fail.

If you encounter an issue with the Debian Bullseye OS image after updating the operating-system (error: filesystem is mounted as read-only), follow our documentation to boot your Instance in recovery mode and recover to a working kernel version.

• fr-par-2
• nl-ams-2
• pl-waw-1 More AZs will follow in the coming weeks.

Joining the ENT1 range, ENT1-XXS and ENT1-XS are now available in all AZ.

The new generation of PLAY2 Development Instances is available in the PAR1, PAR2, AMS2, WAW1, WAW2 Availability Zones, and now also in AMS1.

We also recommend you use the Serverless Functions dashboards in the Scaleway Cockpit to monitor all your metrics and logs and access advanced features (ordering, filtering...)

Key features include:

• AI-optimized with Ampere Altra processors
• Up to 2.8x more energy-efficient than traditional x86
• Configurations from 2-32 vCPUs, 8-128GB RAM

Migrate to COP-ARM for enhanced performance and sustainability. Find out more on our COP-ARM Instances page.

Upon creation, secrets can now be configured with a time-to-live or to expire after a single access through an ephemeral policy. If applied on a secret, all its versions will inherit from the policy.

Refer to the Secret Manager API documentation for more information.

For Organizations that used this permission set, the permission set was renamed from ResellerCustomerBillingManager to BillingManager. Scripts that used the former name of the permission set, such as Terraform scripts, should be updated to use the new BillingManager name.

Llama 3.3 is a fine-tuned version of the Llama 3.1 70b model, which was designed to approach the performance of Llama 3.1 405B on some applications.

To ensure continued compatibility, we recommend upgrading to the latest version (v0.3) as soon as possible.

Refer to our dedicated documentation and learn how to upgrade: Managing Block Storage volumes with Scaleway CSI

Here is their configuration:

• 10 CPU cores
• 10 GPU cores
• 16GB of RAM
• 256GB of storage
• and all the latest innovations from Apple.

For more details and to get started, visit Scaleway's Mac mini M4 product page.

Refer to the auto-generated IAM resources documentation for more information. For each creation or deletion of an auto-generated IAM resource, a log will appear on the IAM logs tab.

Refer to the How to manage blocklists documentation page for more information.

• List-ID
• List-Help
• List-Owner
• List-Post
• List-Subscribe
• List-Unsubscribe
• List-Unsubscribe-Post
• Message-ID

The Standard One Zone storage class comes with a 99% availability SLA, making it ideal for cost-effective storage in production and development environments, media distribution, monitoring, and backup.

Mistral Large 3 is a frontier open-weight multimodal model, optimized for text generation, agentic tasks and image analysis.

• What changed: We refined the underlying calculation used to report memory usage.

• What to expect: You may notice that memory metrics in your dashboards and views appear slightly lower than before. This is an expected result of the improved measurement accuracy, and does not indicate a change in your application's actual performance or resource allocation.

• Scope: This fix applies exclusively to services using sandbox v2.

No action is required on your part.

• What changed: We refined the underlying calculation used to report memory usage.

• What to expect: You may notice that memory metrics in your dashboards and views appear slightly lower than before. This is an expected result of the improved measurement accuracy, and does not indicate a change in your application's actual performance or resource allocation.

• Scope: This fix applies exclusively to services using sandbox v2.

No action is required on your part.

Control plane & add-ons

• Calico: v3.31.3
• Cilium: v1.18.5
• Metrics Server: v0.8.0
• CoreDNS: v1.13.2
• CSI components: v0.3.5
• Cloud Controller Manager (CCM): v0.35.0

Node changelog components

• Linux kernel: 6.8.0-64-generic
• containerd: 2.2.1
• CNI plugins: 1.9.0
• runc: 1.4.0

Feature gates available in 1.35

The following feature gates are available and can be enabled as needed:

• HPAScaleToZero
• CPUManagerPolicyAlphaOptions
• MutatingAdmissionPolicy

For more information, refer to Scaleway's Kubernetes version support policy.

This feature provides quick access to a downloadable PDF summary, supports ESG and Corporate Sustainability Reporting Directive (CSRD) compliance by providing structured environmental data, and allows you to track your infrastructure's environmental performance over time.

The report is available for all active projects and covers the full previous year. Access it now in the Environmental Footprint section of the console.

You can now view your carbon (CO₂e) and water usage trends month over month and instantly identify usage spikes or the results of your optimization efforts via dynamic graphs. No need to download PDF reports anymore to get a high-level view of your impact.

Available now in the Environmental Footprint tab.

Additionally, we have worked on several minor improvements for a better user experience:

• From the Serverless Container list in namespaces, you can now access the logs and metrics of a container
• Creation and last edit dates are now displayed in the overview
• Fixed an issue when deploying Serverless Containers from the Container Registry

• Added creation date to the Job Definition list
• Added creation and last edit dates in the Job Definition overview
• Improved display of mvCPU and memory in the Job Definition list
• Enhanced Job Run list layout for easier ID copying
• Improved container image display and copy functionality

Find out how to activate or configure the operator in the documentation.

• By default, each region will have at least one VPC. You can create more if needed
• Each VPC can have multiple Private Networks, each with its own distinct CIDR blocks
• VPC will now include a DHCP server for each Private Network

• Internet access: allowing external traffic to access the internal resources of a Private Network
• NAT-Gateway: distributing requests to the right IP addresses
• SSH Bastion: allowing and monitoring SSH connections
• Local DNS server: customizing resource names on the Private Network

See our dedicated documentation for full details.

AMS-3 is now fully operational, providing enhanced accessibility and proximity to support our customers in the Benelux region. This comes as the latest development in our commitment to offer an ever-resilient and high-performing cloud infrastructure. Deploy your Instances in AMS-3 now to experience unmatched performance, reliability, and data sovereignty.

Simple and intuitive access management reduces configuration errors and enhances security. Granting granular access to your Object Storage resources has never been easier with our user-friendly interface that allows you to configure both IAM and bucket policies.

Find out how create and manage your bucket policies from the Scaleway console.

To address this issue, Ubuntu 22.04 images used in the Kapsule Kubernetes service have been patched. The patch will be automatically applied to all newly created nodes in managed clusters.

Additionally, the patch will be automatically applied to existing nodes.

In addition to patch versions 1.30.1, 1.30.2, and 1.30.3, releases 1.29.2 to 1.29.6, 1.28.8 to 1.28.11, 1.27.11 to 1.27.15, and 1.26.14 to 1.26.15 are also available.

Read our newest version support policy for more details on the 14-month support timeline.

• application/vnd.openxmlformats-officedocument.spreadsheet
• application/vnd.ms-excel

Find out about the full range of available attachment types in our Transactional Email capabilities and limits documentation page.

All secrets not explicitly tagged appear as opaque, but you can now decide to create new secrets of a specific type (among opaque, username/password, database credentials, SSH key, SSL/TLS certificate and key/value), and secrets versions will be of the same type.

Read our documentation for more details on how to create secret types using the Scaleway API.

• You can now select the version of the OS installed during Mac mini creation.
• Configuration profiles have been updated to allow the management of update policies and the application of beta upgrades.
• The username is now explicitly displayed in the console.

Read our documentation to see the list of actions available. Account actions are tracked in the Paris region.

You can now install a fully functional WordPress website in just one click, no technical knowledge required. Whether you are building a blog, a portfolio, or an online store, getting started has never been easier.

Available now for all Web Hosting plans via the Scaleway console.

They can filter by:

• Client (by customer Organization name)
• Period (by year)
• Category (by product category, e.g. Compute, Storage, Containers, with each category corresponding to a different color in the consumption chart and detailed list)
• Product (by product, with the consumption shown being the sum of all the products of the same type.)

Refer to the documentation page for more details

For Web Hosting users who rely on Redis, we recommend migrating to Scaleway Managed Database for Redis™.

Read more about attaching a Private Network to a Serverless Container.

Read more about attaching a Private Network to a Serverless Function.

• 1.32.[4-7]
• 1.31.[8-11]
• 1.30.[12-14]

We have also added the CPUManagerPolicyAlphaOptions k8s feature gate, and enabled the bpf-lb-external-clusterip cilium option to easily use tools like tailscale or telepresence.

Devstral Small is a fine-tuned version of Mistral Small 3.1, optimized to perform agentic coding tasks. It can be plugged into your favorite IDE.

These new Instances are currently available in:

• fr-par1
• fr-par-2
• nl-ams-1
• nl-ams-2
• pl-waw-2
• fr-par-3 will follow in the coming weeks.

Learn more about POP2, POP2-HC and POP2-HM Instances.

All Load Balancers that were still using SNI routes with HTTP backends on June 1st have now been automatically updated to use the HTTP Host header route instead.

For more information please refer to our product documentation.

• 1.27.2
• 1.26.5
• 1.25.10
• 1.24.14

It now uses the official Debian Unstable cloud-init package.

Users who want to fix their existing Instances may follow this guide.

• The Load Balancer creation process has been streamlined
• The backend management pages have been completely reorganized to provide a clearer and more consistent experience
• You now have access to many backend configuration settings that were previously only available through API

On top of this we have taken the opportunity to introduce 2 new features:

• Backend protection allows limiting the number of sessions on backend servers
• Backend retry logic allows fine tuning how to retry failed connections to backend servers

For more information you can have a look at our updated reference documentation.

Check out our example on sending emails with an SMTP server from Python functions.

With Routed IPs, your worker nodes will receive a flexible IP address, maintaining IP consistency even if the Instance is restarted or the hypervisor is changed.

NAT IPs will reach end of life on October 1st, 2024.

• All information on the migration methods are provided in our documentation about moving Kubernetes nodes to routed IPs.
• Read our blog post to find out more about the changes behind this evolution.

• .art (creation, renewal, transfer): €26.23 excl. taxes
• .at (creation, renewal, transfer): €20.41 excl. taxes
• .it (creation, renewal): €7.81 excl. taxes
• .lu (creation, renewal): €27.78 excl. taxes
• .nl (creation, renewal): €7.71 excl. taxes
• .pl (creation, renewal, transfer): €34.98 excl. taxes
• .tm (creation, renewal, transfer): €558.94 excl. taxes
• .ws (renewal): €30.58 excl. taxes

Please update to DHCP attachment type (recommended) or use an IP address which belongs to the Private Network subnet.

If you already use DHCP attachment mode or static attachment mode with an IP address belonging to the Private Network subnet, you're not affected by this change.

Their lifecycle will now follow the lifecycle for Active status models in our model lifecycle policy.

Devstral is a fine-tune of Mistral Small 3.1, optimized for agentic software engineering tasks. Refer to the official documentation for more information and try Devstral within your IDE.

• Old price: €59.99/month
• New price: €39.99/month

The offer is currently available in PAR-1 and PAR-2, and will soon be available in all regions.

Every new Member must log in using the Member login form and will receive an email to be informed about this change during the migration.

• Old price: €59.99/month
• New price: €39.99/month

Our standard 12-month and 36-month commitments still apply, allowing you to benefit from an additional 5% or 15% discount on your monthly fee, respectively.

With this new pricing, our AMD Ryzen dedicated servers become the market's most affordable 6-core dedicated servers.

Our French Support team has been strengthened and trained to improve your service.

In line with this upgrade, we have refreshed our Support Plans with new pricing, clearer service levels, and expanded options.

The new plans can be found in the Support Plans page of the console.

Find below an overview of the new plans:

• Basic (Free) - 24/7 ticket support, 8h IRT for P1 incidents
• Advanced (€50/month or 5% net spend*) - 2h IRT for P1s, Hotline 7/7 (9AM–6PM)
• Business (€250/month or 10% net spend*) - < 30min IRT for P1s, 24/7 Hotline, Dedicated Premium Team
• Enterprise (€990/month or 20% net spend*) - 15min IRT for P1s, 24/7 Hotline, Premium Team, Technical Account Manager, quarterly reviews, architecture audit, and official certification.

*Higher figure applies.

Refer to the Understanding support plans documentation page for more details about how each plan works.

• SDK:
  • Cockpit
  • Token
  • Grafana User
  • Alerts
• CLI:
  • Cockpit
  • Token
  • Grafana User
• Terraform:
  • Cockpit
  • Token
  • Grafana User

Secret Manager helps you manage the lifecycle of sensitive information:

Secrets are encrypted and stored in a central and secure way, while data access is secured by an authenticated API call to avoid hard-coded transmission and protect your Organization against security breaches.

Routes are used when a Load Balancer has multiple backends, to choose which of them to forward traffic to. There are now two types of routes:

• SNI routes: the only type we previously supported. They use the SNI value provided during the TLS handshake. This route type is suitable for TCP backends when TLS is enabled on the Load Balancer.
• Host routes: this new route type uses the value of the Host header in the incoming HTTP request. It is perfect for HTTP backends when you need to host several services behind your Load Balancer.

The console user interface manages everything for you and configures the route type automatically based on the frontend and backend you use. The API still lets you configure the route type but make sure you update your automation tools, as from June 1st we won't support the use of SNI routes with HTTP backends anymore.

Create a dedicated, private, and flexible Layer 2 networks to which you can attach Scaleway resources such as Instances, Load Balancers, Public Gateways, Elastic Metal, and Managed Databases.

Find more details in our version support policy.

Key features include:

• Nvidia L4 24 GB (Ada Lovelace architecture)
• 4th generation Tensor cores
• 4th generation RT cores (graphics capability)
• Available in multi-GPU 1, 2, 4 or 8 GPUs

Key features include:

• Nvidia H100 80GB (Hopper architecture)
• 4th generation Tensor cores
• Multi-Instance GPU (MIG)
• Transformer Engine
• Available in multi-GPU 1 or 2 GPUs

A setup tutorial will be available soon in your Scaleway console. In the meantime, you can refer to the pgAudit documentation.

For security and compliance, audit logging allows to keep a chronological record of calls made to the Kubernetes API server, investigate suspicious API requests, collect statistics, or create monitoring alerts for unwanted API calls.

All audit logs are centralized in Cockpit for free. Find more information in our dedicated documentation

For seamless deployments, read our documentation:

• how to deploy both x86 and ARM images
• troubleshooting issues with containers not starting on ARM Instances

Read our tutorial to get started with Kubernetes Storage for more information.

Note that:

• Migrations from CSI v0.2 to CSI v0.3 are not yet supported.

• When using the no-scw-csi-v0.3 tag at cluster creation, new clusters can keep using the CSI version 0.2 and its legacy storage class scw-bssd.

Each Serverless Job comes with an ephemeral storage capacity that you can now customize to accommodate more demanding use cases.

Refer to our documentation for more information.

• Simplify tracking of cloud resource costs.
• Access 5 years of data for better analysis.
• Effortlessly monitor costs with customizable filters.
• Use the billing API for seamless data retrieval.

Discover the Cost Manager tool in the billing section of your Scaleway console.

Refer to the How to use the cost manager documentation page for more information.

Clusters still running these versions should be upgraded to a supported minor version as soon as possible. For further details, refer to our version support policy.

Find more details in our version support policy.

Additionally, you now get a breakdown of pricing changes when modifying resources.

Discover all the details in our documentation.

Explore it now in the Scaleway console.

We have now identified and fixed the issue across all regions.

This WAF solution is fully managed and designed to be easy to integrate into existing architectures, particularly when combined with Scaleway's Edge Services.

WAF is currently available only via the Edge Services API.

In addition to patch versions 1.32.0, 1.32.1, 1.32.2, releases 1.31.4, 1.31.5, 1.31.6, and releases 1.30.4 to 1.30.10 are also available. Read our version support policy for more details on the 14-month support timeline.

• Rust 1.85.1 (edition 2024)
• Go 1.23
• Go 1.24
• PHP 8.4
• Python 3.13

Deprecated runtimes

• Python 3.9

End of support

• Rust 1.79
• Go 1.20
• Go 1.21
• Node 18
• Python 3.8

End of life

• Rust 1.65
• Node 14
• Node 16
• Python 3.7

For additional information about Serverless Functions runtimes, check the dedicated documentation.

Members are users who only exist in one Organization and can not own their own Organization. IAM Administrators can manage Member credentials and enforce security requirements for members to ensure Organization security. Inviting Guests remains possible but will be deprecated in the following months.

Read our documentation on Members to learn more.

For more details, check out our documentation for Block Storage and Local Storage.

Key features include:

• Nvidia L40S 48 GB (Ada Lovelace architecture)
• 4th generation Tensor cores
• 4th generation RT cores (graphics capability)
• Transformer Engine
• Available in multi-GPU 1, 2, 4 or 8 GPUs

Read our documentation on the notification manager to learn more.

What's included:

• Track cluster creation, editing, deletion, and configuration changes
• Monitor node pool and node operations
• Log KubeConfig downloads and resets
• Track application deployments and changes

Key benefits:

• Full visibility within 2 seconds of action completion
• Complete audit logs with user IDs, timestamps
• Simple filtering by action type and resource
• One-month log retention

Access this feature in the Audit Trail section of the Scaleway console.

You can now:

• Track IAM principals (User, Application, Group) creation, edition, deletion, and configuration changes
• Monitor changes in permissions through policy changes
• See all actions performed on Members

Find out more about the IAM supported Audit Trail endpoints in the dedicated documentation.

Key benefits

• Full visibility within 2 seconds of action completion
• Complete audit logs with user IDs and timestamps
• Simple filtering by action type and principal
• One-month log retention

Access this feature in the Audit Trail section of the Scaleway console.

You can now track encryption key creation, editing, rotation and deletion. Find out more about the Key Manager supported Audit Trail endpoints in the dedicated documentation.

Key benefits

• Full visibility within 2 seconds of action completion
• Complete audit logs with user IDs, timestamps
• Simple filtering by action type and principal
• One-month log retention

Access this feature in the Audit Trail section of the Scaleway console.

Refer to our dedicated documentation for more information.

This update is designed to support data-intensive workloads requiring increased vertical scaling.

This increase applies to all new and existing 5K and 15K Block Storage volumes across all available regions. You can scale your existing volumes directly through the Scaleway console, via the API, or using the CLI.

For more details on managing your storage, refer to our Block Storage documentation.

Support for the Paris region will be available soon.

This makes it easier to import and manage multiple entries at once.

This makes it easier to import and manage multiple entries at once.

Benefit from a 5% discount with a 12-month commitment, or 15% with a 36-month commitment.

Proxmox VE 9, based on Debian 13 Trixie, brings significant improvements including:

• VM Snapshots on LVM: Support for snapshots on thick-provisioned LVM storage.
• Modernized Web Interface: A revised mobile-friendly interface for easier management on the go.
• Updated Core: Integration of Linux Kernel 6.14, QEMU 10.0, and LXC 6.0.

Proxmox BS 4 also introduces key updates, such as S3-compatible object storage as a backup backend and improved ZFS management.

These operating systems are available for new installations directly from the Scaleway console or the Dedibox console for all compatible hardware configurations.

• 90 days minimum before transitioning objects to Glacier
• 30 days minimum before transitioning objects to Standard One Zone

Existing lifecycle rules below these limits will continue to apply, but we recommend updating existing rules to comply with these new durations for optimal lifecycle management. Any update to an existing lifecycle rule will require compliance with these minimum durations.

Learn more about using your own Grafana with Grafana with Scaleway's datasource in the Using your own Grafana with Cockpit documentation page.

Find out more information on the Kubernetes Private Network documentation

When using this API through Kapsule, Terraform, the SCW CLI or another official open source tool, a tool update is required.

If you encounter any problems with this migration, don't hesitate to contact us on the #load-balancer channel of the Scaleway Community Slack.

Find out more information on the Kosmos documentation.

Key Specifications:

• Processor: AMD EPYC™ 7313P (16 cores, 3.0 GHz)
• Storage: 2 × 1.92 TB SSD NVMe
• RAM: 128 GB
• Bandwidth: 1 Gbps (Public), 10 Gbps (Private)
• Price: See pricing page
• Availability Zones: PAR-1, PAR-2

Designed for professionals, this high-performance server is here to elevate your computing experience. Experience lightning-fast data processing with NVMe storage and enjoy 10 Gbps private bandwidth for seamless communication!

Refer to the How to apply maintenance documentation page for more information.

Choose from over 60 locations worldwide to deploy your Dedibox VPS, ensuring optimal performance and minimal latency for your users. Dedibox VPS offers the power and flexibility of a dedicated server within a shared hosting environment. With Dedibox VPS, you get complete control at a fraction of the cost, making them an ideal solution for individuals, businesses, and developers alike. Enjoy complete control over your VPS, allowing you to install and configure the software to suit your needs.

This price adjustment will subsequently influence the pricing of .FR domains purchased through Scaleway, commencing June 1st, 2024.

• Running complex data processing tasks that require extended execution times
• Hosting long-running webhooks or callbacks that need to handle prolonged requests
• Supporting real-time data streaming applications that require continuous processing
• Enabling wider usage of websockets

• Running complex data processing tasks that require extended execution times
• Hosting long-running webhooks or callbacks that need to handle prolonged requests
• Supporting real-time data streaming applications that require continuous processing
• Enabling wider usage of websockets

• .info: €22.63 excl. taxes
• .mobi: €35.61 excl. taxes
• .studio: €40.53 excl. taxes
• .academy: €43.13 excl. taxes
• .world: €37.93 excl. taxes
• .education: €31.43 excl. taxes
• .media: €41.83 excl. taxes
• .solutions: €28.83 excl. taxes
• .consulting: €47.03 excl. taxes
• .family: €35.33 excl. taxes
• .company: €16.48 excl. taxes
• .events: €39.23 excl. taxes
• .care: €39.23 excl. taxes
• .ninja: €27.53 excl. taxes
• .group: €23.63 excl. taxes
• .network: €31.43 excl. taxes
• .rocks: €19.73 excl. taxes
• .gallery: €24.93 excl. taxes
• .systems: €31.43 excl. taxes
• .business: €16.48 excl. taxes
• .center: €28.83 excl. taxes
• .band: €28.83 excl. taxes
• .plus: €41.83 excl. taxes
• .video: €32.73 excl. taxes
• .international: €28.83 excl. taxes
• .coach: €63.93 excl. taxes
• .games: €30.13 excl. taxes
• .cafe: €41.83 excl. taxes
• .institute: €24.93 excl. taxes
• .legal: €63.93 excl. taxes
• .watch: €39.23 excl. taxes
• .ltd: €27.53 excl. taxes
• .photography: €32.73 excl. taxes
• .house: €40.53 excl. taxes
• .industries: €41.83 excl. taxes
• .show: €39.23 excl. taxes
• .tips: €27.53 excl. taxes
• .vet: €37.93 excl. taxes
• .capital: €63.93 excl. taxes
• .rentals: €40.53 excl. taxes
• .archi: €93.83 excl. taxes
• .cam: €9.33 excl. taxes
• .cyou: €15.83 excl. taxes
• .domains: €39.23 excl. taxes
• .fish: €39.23 excl. taxes
• .healthcare: €78.23 excl. taxes
• .partners: €63.93 excl. taxes
• .beauty: €14.53 excl. taxes
• .dental: €65.23 excl. taxes
• .lgbt: €73.03 excl. taxes
• .market: €41.83 excl. taxes
• .reviews: €56.13 excl. taxes
• .bond: €15.83 excl. taxes
• .place: €19.73 excl. taxes
• .bet: €17.33 excl. taxes
• .furniture: €104.23 excl. taxes
• .recipes: €71.73 excl. taxes
• .audio: €130.23 excl. taxes
• .rodeo: €10.08 excl. taxes
• .be: €7.15 excl. taxes

These updated prices will take effect starting June 17, 2025.

WAF provides automated protection for web applications against common threats and vulnerabilities, including SQL injection, XSS, remote code execution, and more. It is based on the OWASP Core Rule Set to detect and block malicious requests.

M4-M, €0.29/hour. It offers a more powerful alternative to the M4-S, with:

• 10-core CPU
• 10-core GPU
• 32 GB RAM
• 1 TB storage
• 10 Gbps bandwidth

M4-XL, €0.49/hour. The most powerful Mac mini M4 Pro, with:

• 14-core CPU
• 20-core GPU
• 64 GB RAM
• 2 TB storage
• 10 Gbps bandwidth

For more details and to get started, visit Scaleway's Mac mini M4 product page.

HTTP/3 is the next generation of HTTP protocol, built on top of QUIC. HTTP/3 represents a big step forward for internet performance and security. As this is a very new protocol, you have to enable HTTP/3 support if you wish to use it with a Scaleway Load Balancer. HTTP/3 is currently only supported at the frontend.

Find out how to set up your Load Balancers for HTTP/2 and HTTP/3.

Read our documentation to find out how to start using our traces feature.

You can filter on these flags via the API and from the Email activity tab of your Transactional Email domain in the Scaleway console.

For more information, refer to the flags concept page or the documentation on how to manage your email activity.

Read more about Load Balancer logs in our documentation.

Find out more about Secret Manager in our dedicated documentation.

You can check who created, updated or deleted an IAM resource in the dedicated Logs tab on Console, or using the Scaleway CLI.

For more information, refer to the IAM logs documentation.

For more information, refer to the PostgreSQL and FerretDB release notes.

For more information, refer to the pgvector github project.

This allows for seamless retrieval of your sensitive data like API keys, database credentials. This integration reduces complexity while enhancing security by keeping your secrets out of your codebase. Refer to the dedicated documentation for more information.

This rate limit is set at a level that is well above typical usage, so most users should not be impacted.

This rate limit is set above typical usage levels, so most users should not experience any impact on their applications.

This change fixes the creation of job definitions with default values in the console.

This rate limit only applies to queries that require a full DNS resolution, and is enforced independently for each Container instance. This rate limit is set at a level that is well above typical usage, so most users should not be impacted. If you do encounter any issues, reach out to our Support team for assistance.

Refer to the dedicated documentation for more information on DNS query rate limiting.

This rate limit only applies to queries that require a full DNS resolution, and is enforced independently for each Function instance. This rate limit is set at a level that is well above typical usage, so most users should not be impacted. If you encounter any issues, reach out to our Support team for assistance.

Refer to the dedicated documentation for more information on DNS query rate limiting.

Nemotron improves human-like responses in complex tasks, while Molmo provides increased accuracy on multimodal inputs (text and images).

Refer to the dedicated pricing page for more information.

You can find detailed specifications of the configurations below:

The Iridium range is expanding with the addition of 6 new NVMe offerings, from 8 to 64 cores, providing configurations tailored to your needs.

You can find detailed specifications on the configurations below:

In addition to patch versions 1.31.0, 1.31.1, 1.31.2, 1.31.3, releases 1.30.1, 1.30.2, 1.30.3, 1.29.2 to 1.29.6, 1.28.8 to 1.28.11, 1.27.11 to 1.27.15, and 1.26.14 to 1.26.15 are also available. Read our version support policy for more details on the 14-month support timeline.

• improved spam detection
• enhanced analysis powered by machine learning
• greater overall accuracy
• stronger protection against malicious links and attachments

These messages help you quickly understand why a payment failed and provide guidance on how to fix the issue and complete the payment.

Previous authentication method based on JWT will be deprecated gradually. Follow How to migrate from JWT to IAM authentication for private containers.

Find example usage and reference for the new scaleway_datawarehouse_deployment resource in the official documentation.

Previous authentication method based on JWT will be deprecated gradually. Follow How to migrate from JWT to IAM authentication for private functions.

• CPU usage
• Disk read/write
• Network in/out
• Network packets in/out

The managed dashboard gives you an overview on the following metrics:

• Total number of objects
• Total numbers of parts
• Total bucket usage
• Bucket usage per storage class
• Total ingress internal
• Total egress internal
• Total bandwidth internal
• Total ingress external
• Total egress external
• Total bandwidth per type
• HTTP Response Time Distribution
• HTTP Response Time Percentiles

As described in our Certificate Authority update documentation, the migration.iot.fr-par.scw.cloud endpoint is no longer valid.

Key features include:

• Nvidia H100 80GB (Hopper architecture)
• 4th generation Tensor cores
• Multi-Instance GPU (MIG)
• Transformer Engine
• Available in multi-gpu 1 or 2 GPUs

Kubernetes users can now use the scw-loadbalancer-private annotation on their services. Load Balancers are public if this annotation is unset or false. Find the Cloud Controller Manager documentation here

Two options:

• default - controlled isolation: nodes get private endpoints and keep their public IPs, protected behind a security group (dropping inbound traffic)
• full isolation: nodes have only private IPs (Public Gateway required, single egress IP)

The option is immutable after pool creation. Both isolation modes can be used in the same cluster.

The feature is now available in Scaleway console (new node pools), Scaleway API, CLI and Terraform (public_ip_disabled)

Find more details on our reference documentation.

Effective now, every new Kapsule cluster is required to have a (free) Private Network attached.

Starting this deprecation cycle, existing Kapsule clusters must be connected to a Private Network.

The default worker nodes' isolation configuration will keep public IPs for nodes to access the Internet, and add private endpoints for resources to communicate securely. Security groups aren't overridden at migration and RR wildcard DNS still point to public IPs.

More information on the migration steps is provided in our reference documentation.

The service supports 3 protocols: NATS, SQS and SNS (the latter is still in Public Beta). Messaging and Queuing is serverless-based, so you can run it without worrying about provisioning and you only pay for what you use.

The Messaging and Queuing features are available via the Scaleway console, the CLI and Terraform.

Your existing resources will need to be migrated, as communicated to all beta users.

If you have any questions, reach out to the team on the #messaging-queuing community Slack channel.

The service supports 3 protocols: NATS, SQS and SNS (the latter is still in Public Beta). Messaging and Queuing is serverless-based, so you can run it without worrying about provisioning and you only pay for what you use.

The Messaging and Queuing features are available via the Scaleway console, the CLI and Terraform.

Your existing resources will need to be migrated, as communicated to all beta users.

If you have any questions, reach out to the team on the #messaging-queuing community Slack channel.

The service supports 3 protocols: NATS, SQS and SNS (the latter is still in Public Beta). Messaging and Queuing is serverless-based, so you can run it without worrying about provisioning and you only pay for what you use.

The Messaging and Queuing features are available via the Scaleway console, the CLI and Terraform.

Your existing resources will need to be migrated, as communicated to all beta users.

If you have any questions, reach out to the team on the #messaging-queuing community Slack channel.

Scaleway's Console, API, CLI, and Terraform now all support multi-AZ Kubernetes Kapsule clusters.

Instructions are available in our reference documentation. Step-by-step guidance also provided in our complete Terraform tutorial.

Pre-requisite: Kubernetes Kapsule cluster connection to a Private Network.

See our product documentation for more information.

For more information, refer to the Subscriptions documentation.

For more information, refer to the Subscriptions documentation.

Parameters tools and tool_choice of our OpenAI-compatible chat API are now accepted for models with this capacity.

Read our dedicated documentation and tutorial to get started!

Find example usage and reference for the new scaleway_inference_deployment resource in the official documentation.

• InterLink Hosted: connect your on-premises infrastructure to your Scaleway VPC via a partner, with a mutualized third-party interconnection
• InterLink Self-Hosted: directly connect your on-premises infrastructure to your Scaleway VPC with your own physical dedicated connection

Check the full list of supported file extensions in the supported file types documentation.

• CPU: Intel Xeon E3 1220 or equivalent - 4C / 4T - 3.1 GHz
• RAM: 32 GB
• Disks: 2 × 1.02 TB SSD
• Public bandwidth: Up to 1 Gbps
• Price: €27,99/month

Check the full list of supported file extensions in the supported file types documentation.

Following our model lifecycle policy, this model will remain accessible until November 14th, 2025. After this date, requests will be routed automatically to a similar model: Qwen3 Coder.

• If the Job definition uses an image from Container Registry, you must now add, at minimum, the ContainerRegistryReadOnly permission.
• If the Job Definition consumes data from Secret Manager; you must now add, at minimum, the SecretManagerSecretAccess permission.

Price drops

The monthly prices for the following offers have been reduced:

Configuration upgrades (no price change)

To provide you with more storage capacity at no extra cost, the SSD disk capacity has been doubled on the offers below:

Note: The other specifications (CPU, RAM) for these offers remain unchanged.

Both Whisper Large v3 and Voxtral Small 24B can now be queried to perform audio transcription using Audio Transcriptions API.

Patch versions 1.27.3, 1.27.4, 1.26.6, 1.26.7, 1.25.11, 1.25.12, 1.24.15, 1.24.16 have also been released.

Read our version support policy for more information.

Dedicated control planes are available in the Scaleway console, the CLI, and Terraform. Refer to our documentation for more information.

Access logs are invaluable tools that record detailed information about every interaction with your objects in the Object Storage service. These logs include essential data such as who accessed your objects, when they accessed them, and from which IP address. With access logs enabled, you can gain deeper insights into how your data is being used, and enhance your security and compliance efforts.

While the feature will no longer be available from the Scaleway console, it is still available in the API, though entering a deprecation cycle.

For more information, refer to the How to create Read Replicas documentation page.

From Monday 9th September 2024, the pricing for .com and .net TLDs will also change.

Refer to the dedicated page to consult our new pricing.

For more details, please refer to our dedicated documentation.

Released July 2024, these models all support a very large context window of up to 128k tokens, particularly useful for RAG applications.

• Network I/O Monitoring: Gain insights into network traffic and usage.
• Key Hit/Miss Ratio: Track the ratio of cache hits to misses for improved cache performance.
• Client Metrics:
  • Monitor the number of active clients.
  • Track blocked clients (e.g., during BLPOP operations).
• Key Expiration Stats:
  • Ratio of keys approaching expiration.
  • Ratio of keys that have expired.
• Key Eviction Ratio: Monitor the ratio of keys evicted due to memory pressure.

Refer to our documentation for code examples, including usage with Pydantic models or manual schema definitions.

The minimum size is 1 GB, and the maximum size depends on your RAM, to an upper limit of 24 GB.

This is now the default sandbox version for new Functions on the console, as it offers enhanced performance.

This is now the default sandbox version for new Functions on the console, as it offers enhanced performance.

With a privacy-focused, fully managed stack, this solution opens up new possibilities for applications that require both textual and visual understanding, knowing your images remain safe.

Read our dedicated documentation to get started with Pixtral.

• EM-B220E-NVME
• EM-B320E-NVME
• EM-B420E-NVME
• EM-B520E-NVME
• EM-I115E-NVME
• EM-L220E-NVME
• EM-L420E-NVME

To activate it, go to the Elastic Metal server section of the Scaleway console and click Settings > Public bandwidth > Manage bandwidth.

Prices of this option are available here.

For users who still reach the limit (429 error code), the process for asking for a rate limit upgrade remains unchanged, through opening a support ticket.

Read our dedicated documentation for more information and to purchase your first Savings Plan.

• 1.33.0 – 1.33.4
• 1.32.8
• 1.31.12

Kubernetes 1.30 updates

• Calico: v3.28.x → v3.28.5
• Cilium: v1.15.x → v1.15.19
• Metrics Server: v0.7.x → v0.7.2
• CoreDNS: v1.11.x → v1.11.4
• CSI: v0.3.x → v0.3.5
• CCM: v0.32.0 → v0.34.0

Node components:

• Kernel: 6.2.0-36-generic
• Containerd: 1.7.28
• CNI: 1.7.1
• Runc: 1.3.0

Kubernetes 1.31 updates

• Calico: v3.28.x → v3.28.5
• Cilium: v1.16.x → v1.16.12
• CoreDNS: v1.11.x → v1.11.4
• CSI: v0.3.x → v0.3.5
• CCM: v0.32.0 → v0.34.0

Node components:

• Kernel: 6.8.0-49-generic
• Containerd: 1.7.28
• CNI: 1.7.1
• Runc: 1.3.0

Kubernetes 1.32 updates

• Calico: v3.29.x → v3.29.5
• Cilium: v1.16.x → v1.16.12
• CoreDNS: v1.12.x → v1.12.2
• CSI: v0.3.x → v0.3.5
• CCM: v0.32.0 → v0.34.0

Node components:

• Kernel: 6.8.0-49-generic
• Containerd: 1.7.28
• CNI: 1.7.1
• Runc: 1.3.0

Kubernetes 1.33 updates

• Calico: v3.30.2
• Cilium: v1.17.6
• Metrics Server: v0.8.0
• CoreDNS: v1.12.2
• CSI: v0.3.5
• CCM: v0.34.0

Node components:

• Kernel: 6.8.0-49-generic
• Containerd: 2.1.3
• CNI: 1.7.1
• Runc: 1.3.0

Available Feature Gates in 1.33:

• HPAScaleToZero
• DRAAdminAccess
• DynamicResourceAllocation
• PodLevelResources
• CPUManagerPolicyAlphaOptions

Try them now to benefit from improved token generation speed (up to x2 on most models).

This makes it easier to import and manage multiple entries at once.

• Bulk order: You can now place multiple orders for the same Elastic Metal resource in a single API request.
• Audit Trail: Elastic Metal activity can now be tracked in the Audit Trail product, directly in the console. All API gateway events are listed. Find out which endpoints are supported in the dedicated documentation.

Voxtral is a frontier chat and audio model that can transcribe or understand audio files using Chat Completions API.

Get started and read the documentation.

• Calico: v3.30.3
• Cilium: v1.18.1
• Metrics Server: v0.8.0
• CoreDNS: v1.12.3
• CSI: v0.3.5
• CCM: v0.34.0

Node components:

• Kernel: 6.8.0-64-generic
• Containerd: 2.1.4
• CNI: 1.7.1
• Runc: 1.3.1

Available Feature Gates in 1.34:

• HPAScaleToZero
• CPUManagerPolicyAlphaOptions
• ImageVolume

New cilium_native CNI

On Kapsule 1.34, cilium_native is available as a new CNI. This is cilium with a custom configuration that enables its native mode. For more information, please read this cilium documentation.