Hosting, regions and deployment status
RapidValue is a Scaleway-hosted managed SaaS for Europe. Amsterdam
(nl-ams) is the primary region, on European-owned cloud infrastructure with
customer data hosted in the selected EU region. EU hosting is the standard
European offer; it is not a premium hosting tier.
The production environment has run on Scaleway Amsterdam since 24 September 2026 (staging since 18 September). The earlier AWS-based runtime has been stopped. That is a statement about where the platform runs, not a claim that it is generally available: no customer is live on it yet, no service-level agreement is in force, and the items under "Not yet in place" below are the gate before the first customer goes live.
Status at a glance
| Deployment model | Intended use | Status | Availability and support |
|---|---|---|---|
| RapidValue-managed SaaS on Scaleway Amsterdam | Default for European customers | Production environment running; no customer live yet | Not yet generally available; no contractual service level |
| RapidValue-managed SaaS on AWS | Non-EU regions or an explicit customer requirement | Not currently operated | The AWS runtime was stopped in September 2026; a non-EU region would be scoped separately |
| Customer-managed cloud | Enterprise customers requiring deployment in their own cloud account | Future enterprise option | Separate pricing, implementation and support model; not a standard feature today |
| Customer-managed on-premises control plane | Enterprise customers requiring the whole platform in their own environment | Future enterprise option | Separate pricing, implementation and support model; not a standard feature today |
| Tier-3 agent in a customer network | Local connector reach and local credential custody while using managed SaaS | Supported hybrid component | This is not the same as a customer-hosted control plane; see Tier-3 Hybrid Architecture |
Customers principally choose a data region and hosting requirements. The underlying provider is an implementation choice unless a contract or technical requirement makes it material. The application architecture remains provider-neutral, with provider adapters for the services that cannot be made portable.
What is in place
The production topology on Scaleway Amsterdam consists of two application nodes in separate availability zones behind a load balancer, a highly available managed PostgreSQL 16 database on a private network, managed object storage and a private container registry. Beyond the topology:
- Deployments release an immutable image pair by digest, node by node, behind a health gate. Production is promoted only from a release that passed staging.
- Migrations and tenant isolation: schema migrations run as a separate step before a release, the application connects with a restricted database role, and a row-level-security canary is part of the platform health check.
- Monitoring: logs and metrics are collected centrally, managed alerts and an on-node monitor watch the platform, and an external health check runs every fifteen minutes, including certificate expiry. Public endpoints use managed HTTPS with certificate renewal.
- Backup and restore: the database keeps daily automatic snapshots (30 days in production); per-tenant backups are written to object storage; and since 22 October 2026 an encrypted copy of the database is written to another Scaleway region (Paris) every day. A database restore from a production snapshot was exercised on 4 October 2026, and a full rebuild in the other region from that daily copy on 23 October 2026, both with a recorded result. Recovery targets (how much data and how much time a recovery may cost) are proposals, not commitments.
Not yet in place
Each item below is a gate before the first customer goes live. Passing one does not imply the others.
- Rollback exercised in production. A rollback workflow exists, is documented and was drilled on staging on 15 October 2026; a recorded production rollback run does not exist yet.
- A complete failure-mode validation of the whole topology (node loss, zone loss, database failover) with recorded results. Region loss has been drilled (see above); the rest has not.
- Someone to page. Alerts reach the operating team's mailbox and the repository, but there is no staffed on-call rotation. The public status page shows component status and incidents.
- A service level. The intended commitment is 99.5% availability for every customer, with 99.9% as the goal; neither is contractual today.
- The first-customer gate set on 6 October 2026: a completed external penetration test with findings remediated, a data processing agreement with a sub-processor list, and a customer-ready incident runbook and support inbox.
Security and privacy boundaries
Hosting on Scaleway supports two concrete statements: the infrastructure provider is European-owned, and customer data can be hosted in a selected EU region. It does not by itself prove compliance with a customer's regulatory framework, create immunity from foreign law, or remove the need to review contracts, subprocessors, support access and data flows.
In every managed deployment, the synced governance model—identities, accounts, entitlements, grants, workflow and audit data—lives in the selected control-plane region. A Tier-3 agent changes connector reach and credential custody; it does not keep the synced governance model out of the control plane. See Tenant isolation for the application and database boundary.
Provider-specific operations
Generic product documentation describes outcomes: health checks, migrations, backups, restores, monitoring and rollback. Operator runbooks must name their provider explicitly because the commands and failure modes differ.
- AWS RDS, Secrets Manager, KMS, IAM and Route 53 instructions are AWS runbooks, not generic product requirements.
- Scaleway Database, Secret Manager, Object Storage, IAM and DNS instructions are the production runbooks; AWS runbooks describe a retired runtime.
- A provider adapter is supported only when its operational path is tested; API similarity alone is not enough.
Commercial and support boundary
European managed SaaS does not carry a premium solely because it uses the European default hosting target. An explicit non-default region, customer-cloud deployment or on-premises control plane can materially change implementation, operations and support responsibilities and therefore needs a separate scope.
Do not plan a customer-cloud or on-premises rollout from this page. Those models become orderable only when RapidValue publishes a supported topology, delivery process, responsibility matrix and commercial terms for them.
Further reading:
- Tier-3 Hybrid Architecture — connector execution and credential custody
- Tenant isolation — the tenant boundary inside a shared control plane
- Environments and staged config — application configuration across dev, acceptance and production
- What we don't claim — limits that are deliberately public