PlanetScale for Postgres
PlanetScale for Postgres provides fully-managed, high availability PostgreSQL database clusters built on modern cloud infrastructure. With PlanetScale Metal, you get blazing-fast Postgres, reliability, and the developer experience PlanetScale is known for, in both AWS and GCP.
Why choose PlanetScale for Postgres?
Extreme performance
- Ultra-low latency queries
- PlanetScale Postgres on Metal outperforms Aurora, AlloyDB, Neon, Heroku, and Supabase in benchmarks
High availability
- Primary + 2 replicas across 3 availability zones by default
- Automated failovers with our custom operator
Database branching
- Create isolated database environments from production backups
- Cost-optimized single-instance branches for development
Query insights
- Automatic detection of performance anomalies
- Query execution plan analysis and recommendations
Flexible configuration
- PlanetScale Metal or network-attached storage options
- Curated extensions and custom parameter tuning
Simple migrations
- Guided zero-downtime migration tools for all providers
- Hands-on support from our team available
Zero-downtime migration. No maintenance windows. No frantic debugging. P99 latency dropped to consistently 1.0ms, and API response times improved from 550ms to 300ms.
— Anthony Ronning, CTO @ Maple AI
Why We Migrated from Neon to PlanetScale
QPS jumped from 20 to 1,000 (50x improvement!) and costs dropped from $900/mo to $90/mo (90% reduction!). We're excited that we no longer have to worry about handling scale as we grow.
— Supermemory Team, Founder @ Supermemory
Supermemory just got faster on PlanetScale
Immediately upon hot-swapping the environment variables, our P95 query time dropped significantly across the board. The most drastic improvement was with our pgvector workload. We saw a 91% decrease in latency.
— Jake Casto, Founder @ Layers
Scaling Real-time Discovery: Inside Layers' PlanetScale Migration
How does PlanetScale for Postgres work?
PlanetScale for Postgres is built on a foundation of modern cloud infrastructure, combining the reliability of Postgres with PlanetScale's expertise in running databases at scale. Our architecture is designed for extreme fault tolerance and performance.
High availability architecture
When you create a Postgres database on PlanetScale, you automatically get a highly-available cluster distributed across three availability zones. The cluster consists of one primary instance two replica instances for high availability.
Our custom Kubernetes operator monitors the health of all instances. If the primary becomes unavailable due to a node or drive failure, the operator automatically promotes one of the replicas to become the new primary. In most cases, failover completes in under 30 seconds.
Configuration flexibility
You can choose between different CPU architectures and storage types to optimize for your workload:
CPU Architecture: Select ARM64 (AWS Graviton) for cost-optimized performance, or x86-64 (Intel/AMD) for maximum single-threaded performance.
Storage: PlanetScale Metal provides direct-attached NVMe storage for maximum IOPS and minimum latency. Network-attached storage (EBS) offers flexible scaling up with configurable IOPS.
Observability and insights
Every Postgres database includes comprehensive observability tools built-in. Query Insights automatically captures slow queries, shows execution plans, and provides schema optimization recommendations. Metrics dashboards display CPU, memory, storage, and connection statistics in real-time. All logs are centralized with 7-day retention and full-text search capabilities.
PlanetScale for Postgres resources
Quickstart guide
Deploy your first PostgreSQL database on PlanetScale and learn the basics in under 10 minutes.
Official documentation
Explore comprehensive documentation covering architecture, configuration, monitoring, migrations, and best practices.
Benchmark results
See how PlanetScale for Postgres compares to Aurora, AlloyDB, Neon, Supabase, and other providers.
Migration guides
Migrate from AWS RDS, Google Cloud SQL, AlloyDB, Neon, Supabase, or any PostgreSQL provider.
Customer stories
Read how companies like Convex, Supermemory, and Layers migrated to PlanetScale and improved performance.