[espadrine]

The way I see it, Scaleway vs. AWS is the old tradeoff: enterprise-grade hardware vs. commodity hardware with resilient software.

You can set up one AWS EC2 backed by an RDS.

Or you can set up two Scaleway servers in failover, backed by three CockroachDB Scaleway nodes.

In the second approach, you need more cheap things, but the total price can still be lower. On the other hand, you should not expect the same reliability from a single Scaleway node, compared to a single EC2.

[luhn]

EC2 isn't enterprise hardware, it's all commodity and, more lately, custom-designed silicon. It certainly fails, and Amazon's own recommendations call for redundancy using things like EC2 auto-scaling groups and RDS multi-AZ.

EC2 has a slight leg up because EBS stores data redundantly, so a single disk failure won't knock out your data. But failure is still possible, and a durable storage strategy should involve backups.