What it does
Scaleway Serverless Jobs runs container-based tasks that finish and stop, rather than keeping an HTTP service online. A job definition specifies the image, command, region, CPU and memory; runs can be started manually or scheduled, and multiple tasks can be parallelized. Scaleway manages the execution infrastructure, while users bring a container image and decide where persistent inputs and outputs should live. It suits periodic data processing, migrations, media conversion and other background work that does not need a continuously listening server.
A batch runner, not a web service
Jobs are intended for commands that run to completion. The product page describes scheduling, queueing and parallel batch processing; it explicitly contrasts Jobs with request-driven Serverless Containers. This makes it a fit for a nightly import or a media conversion task, not for serving a website or API endpoint.
What a first run needs
The documented quickstart requires a Scaleway account with suitable organization permissions and a container image already pushed to a Container Registry namespace. In the Console, create a job definition, choose the image and tag, region, and resource allocation, then set a startup command. After saving the definition, open it and choose Run job; the execution appears under Job runs, where its logs can be opened. Those are documented steps, not actions performed for this listing.
For recurring work, add a schedule or invoke the job through Scaleway's API or CLI. The service also integrates with Scaleway storage and databases, which gives a stateless run somewhere to read inputs from and persist its results.
Cost and operating boundaries
Billing is pay-as-you-go for allocated CPU and memory while a run executes, with a monthly free allowance measured in vCPU-seconds and GB-seconds. Local storage is ephemeral, there is no SSH access, and the FAQ says VPC and Private Network support is not yet available; outbound IPs are not dedicated or predictable. Plan external connectivity and durable state around those limits.
Strengths
- Scaleway manages the job execution backend, and definitions can be run manually, on a schedule, or in parallel.
- CPU- and memory-based billing has no idle-capacity charge, and the FAQ documents a monthly free allowance.
- The product exposes API and CLI workflows for automation and can integrate with storage and database services.
Watch out for
- A run needs a container image in a registry; the quickstart requires a pushed image and recommends Scaleway Container Registry over external public registries because of their uncontrolled rate limits.
- Runs are stateless with ephemeral local storage, provide no SSH access, and do not currently support VPC or Private Networks or predictable dedicated IPs.
- The quickstart says Jobs do not provide an interactive web interface or automatic scaling of job runs; schedule and execution behavior must be configured for the workload.
Best for
- Scheduled data imports, transformations, and log-processing batches
- Background media conversion or machine-learning jobs that run to completion
- Automated migrations and maintenance scripts that need containerized execution rather than a persistent web server
Sources
- Serverless Jobs | Scaleway (Oct 26, 2026)
- Jobs - Quickstart | Scaleway Documentation (Oct 26, 2026)
- Serverless Jobs FAQ | Scaleway Documentation (Oct 26, 2026)
- Serverless Pricing | Scaleway (Oct 25, 2026)