Managed Relational Database

What each cloud calls it

Features

• OVHcloud, Scaleway, STACKIT, and IONOS all offer managed PostgreSQL. MySQL and MariaDB availability varies: Scaleway supports both PostgreSQL and MySQL; OVHcloud Managed Databases covers PostgreSQL, MySQL, and MariaDB among other engines; STACKIT PostgreSQL Flex focuses on PostgreSQL only; IONOS DBaaS covers PostgreSQL. If your workload runs MySQL or MariaDB, verify engine support before selecting a provider.
• High-availability configurations (standby replica with automatic failover) are available on all four European offerings. The failover time and RPO/RTO guarantees differ. Scaleway and OVHcloud publish SLAs for managed database availability; check current SLA terms as they are updated periodically.
• Read replicas for horizontal read scaling are supported on OVHcloud and Scaleway. STACKIT PostgreSQL Flex is designed around single primary plus standby HA rather than a read replica fleet. IONOS DBaaS supports read replicas on PostgreSQL.
• Maximum instance sizes differ. AWS RDS supports instance types up to 96 vCPU and 768 GB RAM. European managed database offerings cap out at smaller configurations. For very large single-node databases, check current instance tier limits before assuming a direct migration path.
• Managed extensions: RDS and Cloud SQL support a broad set of PostgreSQL extensions. European providers support common extensions (PostGIS, pg_trgm, uuid-ossp, pgvector) but the supported extension list is narrower. Verify your required extensions are available before committing.

Pricing

• Pricing models differ. RDS charges for instance hours, storage (per GB-month), I/O (for io1/io2 volumes), and data transfer. European providers typically bundle storage with the instance tier or charge separately per GB without I/O surcharges.
• Multi-AZ RDS roughly doubles the instance cost. European HA configurations (standby replica) are similarly priced at approximately 2x the single-node cost. The relative cost of HA is comparable across providers.
• Automated backups are included on all European managed database offerings within a retention window (typically 7 to 30 days). Longer retention requires manual backup to object storage. RDS charges separately for backup storage above 100% of the database size.
• Data egress from the database is significantly cheaper on European providers. For applications that run analytics against production data or replicate to external systems, the egress cost difference can be material.

Tooling and ecosystem

• Connection is via standard PostgreSQL wire protocol. Application-level connection strings change only the host endpoint. No client library changes are required.
• Terraform providers for OVHcloud, Scaleway, STACKIT, and IONOS all cover managed database creation, user management, and backup configuration. Parameter/configuration management is more limited than the RDS parameter group system; advanced PostgreSQL configuration (wal_level, max_connections, shared_buffers) is typically set via the provider console or API, not all parameters are exposed.
• RDS Proxy (connection pooling) has no direct managed equivalent on European providers. PgBouncer deployed as a sidecar or a small separate VM is the standard replacement. Scaleway and OVHcloud have explored managed connection pooling; check current product availability.
• Database migration tools (AWS DMS, pglogical, pg_dump/pg_restore) work cross-provider. For live migrations with minimal downtime, pglogical replication from RDS to a European managed PostgreSQL is a common pattern.

Compliance

• All four European providers store database data within EU datacentres. OVHcloud and Scaleway carry SecNumCloud in addition to ISO 27001 and SOC 2. STACKIT and IONOS hold ISO 27001, SOC 2, and BSI C5.
• Encryption at rest is on by default using provider-managed keys. Bring-your-own-key options vary: Scaleway integrates with its Key Manager for volume-level encryption; OVHcloud and STACKIT offer KMS integration for managed databases; IONOS relies on provider-managed encryption.
• Audit logging (who connected, which queries ran) is available on all providers but the implementation and log routing differs. RDS exports logs to CloudWatch; European providers ship logs to their observability platforms (Scaleway Cockpit, OVHcloud Logs Data Platform). Configure log shipping to your own storage for long-term retention.
• GDPR right-to-erasure workflows: all providers support point-in-time restores and backup deletion, but coordinating data deletion across backups and replicas remains a customer responsibility regardless of provider.

Migration considerations

• Schema and data migration. pg_dump and pg_restore are the most reliable path for offline migrations. For databases under ~100 GB, a maintenance window migration is practical. For larger databases or strict downtime requirements, set up logical replication (pglogical or pg_basebackup with WAL shipping) before the cutover.
• RDS-specific features with no equivalent. RDS Performance Insights, Enhanced Monitoring, and Aurora-specific features (Aurora Serverless, Aurora Global Database) have no direct equivalent on European managed PostgreSQL. Replace Performance Insights with provider-native metrics and a tool like pgBadger or pg_activity for query analysis.
• Parameter group differences. RDS parameter groups expose hundreds of PostgreSQL configuration parameters. European providers expose a subset. If your workload relies on non-default values for parameters like work_mem, max_parallel_workers, or checkpoint_completion_target, verify they can be set on the target provider.
• Automated minor version upgrades. RDS can apply minor version upgrades automatically during a maintenance window. European providers vary in their minor upgrade automation; some require manual initiation. Factor this into your operational runbook.
• Connection string updates. RDS endpoints follow an AWS-specific pattern. Update all application configuration, secrets, and connection pooler config to point at the new endpoint. Use a DNS alias or environment variable abstraction to simplify the cutover.

Further reading

• Why European companies are migrating away from US cloud giants→