Runs in your browser · nothing uploaded
What would your Azure bill cost on Scaleway?
Net of the engineer time to run it. Drop in one month of your Azure cost export and get every instance, volume, bucket and managed service priced on Scaleway list prices, every service graded by effort, and a one-off cost with a payback month.
- Every server matched to a real type
- Standby, backups and engineer hours included
- A likely range, not a single number
What moves, and what it becomes on Scaleway
The services a typical Azure bill is made of, graded by effort: 4 Easy, 10 Moderate, 2 Hard. On a real report each line carries your own cost and the ratio or usage model behind the figure.
| Service | Effort | On Scaleway |
|---|---|---|
| Virtual Machines | Easy | Scaleway Instances or Scaleway Elastic Metal servers |
| Storage | Easy | Scaleway Block Storage for disks, Scaleway Object Storage for blobs, NFS / SMB on Scaleway Instances for Files |
| Azure Database for PostgreSQL | Moderate | PostgreSQL on Scaleway Instances (Patroni or replicas, pgBackRest), or a Scaleway managed database |
| Azure SQL Database | Hard | PostgreSQL with a migration, or SQL Server on a Scaleway VM (licence extra) |
| AKS | Moderate | Scaleway Kubernetes Kapsule |
| App Service | Moderate | Containers on Scaleway with Coolify / Dokploy |
| Azure Functions | Moderate | Scaleway Serverless Functions / Containers, or long-running services on Scaleway Instances |
| Cosmos DB | Hard | ScyllaDB or PostgreSQL with a data-model rewrite, or a Scaleway managed database |
| Azure Cache for Redis | Moderate | Valkey on Scaleway Instances, or a Scaleway managed database |
| Bandwidth | Easy | Scaleway included traffic |
| Load Balancer | Easy | Scaleway Load Balancer or HAProxy / Traefik |
| Application Gateway | Moderate | Scaleway Load Balancer + Traefik / NGINX on Scaleway Instances; Coraza for WAF rules |
| Log Analytics | Moderate | Loki + Grafana |
| Event Hubs | Moderate | Redpanda on Scaleway Instances, or a Scaleway managed database |
| Azure OpenAI | Moderate | OpenAI / Anthropic APIs directly, or open models on GPU Scaleway Elastic Metal servers |
| Databricks | Moderate | Spark on Scaleway Elastic Metal servers, or Databricks on your own cloud |
Read the Azure worked example for a full file taken end to end. Easy: a drop-in replacement. Moderate: the same software, self-hosted, so you patch, back up and fix it. Hard: no equivalent; redesign it, buy a SaaS, or leave it where it is. The full map, with every ratio and its derivation, is on the model page.
About Scaleway
France (Paris), Netherlands (Amsterdam), Poland (Warsaw)
French and EU-only, with the broadest managed catalogue of the European destinations: managed Kubernetes and databases, serverless, object storage and GPU instances. Instances include outbound traffic up to a cap per size. Choose Scaleway when you want a managed layer under EU jurisdiction and your users are in Europe; it costs more than Hetzner per server, and object storage is priced closer to the hyperscalers.
Scaleway list prices, 18 Oct 2026
- Smallest cloud server (2 vCPU, 2 GB)
- €6.56/mo
- Smallest dedicated server (12 vCPU, 64 GB)
- €134.99/mo
- Block storage
- €0.09 per GB-month
- Object storage
- €0 for 0 TB
- Outbound traffic
- unmetered
EUR excl. VAT, converted at 1.17 USD in the calculator. Full price list and source.
How to get your Azure export
- 1Create a Cost Management export (FOCUS or amortized) and download last month.
- 2Drop it on the calculator.
- 3Read the net figure and change any assumption.
The exact clicks, console links and a sample file to compare against are under “Where do I get this file?” on the calculator.
Questions we get asked
- What file do I upload?
- One month of a cost export: an AWS Cost and Usage Report (CUR 2.0 or legacy, CSV or Parquet) or a Cost Explorer CSV, a Google Cloud BigQuery billing export saved as CSV, or an Azure Cost Management export. “Where do I get this file?” next to the upload box has the exact clicks for each.
- Is my data uploaded?
- No. The file is read inside this browser tab by DuckDB-WASM and never leaves it; you can watch the network tab while it runs. The only thing we ever receive is the address you type if you email yourself the link, plus anonymous counts of what was used.
- How accurate is this?
- Servers, storage and traffic are priced from your actual usage at list prices, so that part is as good as the price list. Managed services without a usage model (queues, search, analytics) are priced by a ratio from our service map and carry ±50%. Engineer time is an assumption you can edit. The report shows what share of the figure is priced from usage and gives a likely range rather than a single number. A Cost Explorer CSV has service totals only, so hosts are estimated from spend, roughly ±25%.
- Can I move only part of it?
- Yes, and that is often the right first step. Storage and egress-heavy workloads move easily and save the most per hour of effort; a managed database or a serverless stack can stay where it is. Filter the services table by grade to see what moves cleanly.
- What about compliance and latency?
- Every destination publishes GDPR documentation and most hold ISO 27001; SOC 2 reports and HIPAA agreements are rarer, so check that provider’s trust page before planning. On latency: for European users, moving from an EU hyperscaler region to Falkenstein, Gravelines, Paris or Amsterdam changes little; for US users choose a US location. A CDN in front stays as it is; the report prices it separately.
Your Azure bill on Scaleway, in under a minute.
One month of a cost export is enough. It is read in your browser and never uploaded; only the number comes out. If the honest answer is “stay”, the report says so.