Managed databases built for continuous uptime.
Managed MySQL clusters with automated failover, point-in-time recovery and capacity that follows your load — reachable privately from your Joy servers over the backbone, backed up to a second region, and operated by the team that runs the platform.
A database is the part of a stack that cannot simply be redeployed when it breaks. Joy Database takes the operational work off your servers: a primary with a synchronous replica on separate Compute Fabric nodes, automatic failover when a node is lost, continuous binary-log shipping so you can restore to any second in the retention window, and monitoring by the same on-call engineers who look after the hypervisors.
Clusters are created in the region where your application runs and are reached over the private backbone, so traffic never crosses the public internet and latency from a Joy VPS in the same region is sub-millisecond. Storage sits on NVMe; capacity — vCPU, memory and disk — is resized live as your dataset and traffic grow, and can be scheduled to follow daily peaks. MySQL is first; PostgreSQL and Redis follow on the same control plane.
Database is in build. If you need a managed database on Joy today, the Enterprise team runs MySQL and PostgreSQL for customers under a managed-operations contract on dedicated servers or VPS; open a ticket to join the early-access list for the self-service product.
D1 · D2 · D3
The three things this product line is built around — the same three boxes you see on the tile.
High-performance data infrastructure
NVMe storage, host-passthrough CPU and private backbone connectivity give real-time access from your application servers; connection pooling and read replicas take load off the primary.
Built for critical data
A synchronous replica on a separate node, automatic failover, daily backups replicated to a second region and point-in-time recovery — designed for continuous uptime rather than best effort.
Managed operations
Deploy MySQL today and PostgreSQL or Redis as they open, with backup automation, replication, version upgrades and failover tested by the team that operates the cluster.
Specification
| Engines | MySQL 8 (first), PostgreSQL and Redis on the same control plane |
|---|---|
| Topology | Primary + synchronous replica on separate Compute Fabric nodes; optional read replicas |
| Failover | Automatic promotion when the primary node is lost; the connection endpoint does not change |
| Backups | Daily full backups plus continuous binary-log shipping; point-in-time restore within the retention window; copies in a second region |
| Storage and compute | NVMe; vCPU, memory and disk resized live; scheduled scaling for known peaks |
| Connectivity | Private backbone from Joy servers in the same region; optional public endpoint with allow-lists and TLS |
| Security | TLS in transit, encrypted backups, per-database users, audit of staff access visible to you |
| Monitoring | Query and replication metrics in the console; alerts by email and ticket |
Three steps
Create a cluster
Choose the engine, region and size. The primary and replica are provisioned on separate nodes and the private endpoint appears in the console.
Connect privately
Use the private hostname from your Joy servers in the same region, or enable a public endpoint with an allow-list and TLS.
Scale and recover
Resize live as you grow, add read replicas, and restore to any point in the retention window from the console or API.
Where Database fits.
Questions about Database.
When can I order?
Can I reach the database from outside Joy?
Do I still need my own backups?
Which engines are supported?
More answers in the full FAQ — billing, network, DDoS, backups, support and legal.
Sized per cluster, billed monthly
Pricing is published at general availability and will follow the per-component model used for custom VPS builds. Managed databases under Enterprise contracts are quoted per case today.
Works with
Build Database into your stack.
Create an account and deploy in under ninety seconds, or talk to sales about a design for your workload — regions, sizes, timelines and a quote within one business day.