PlanetScale PS-5 Single-Node vs Larger Cluster — when does one get cheaper?
Choose whether new workloads fit PlanetScale's PS-5 single-node option or need higher-availability cluster configurations after the lower-cost tier launched.
Blockers
- package/planetscale-postgres-ps-5-single-node incompatible with package/planetscale-postgres-metal
Who this is for
- cost-sensitive
- low-ops
- small-team
Candidates
PlanetScale Postgres PS-5 single node
As of 2026-04-08, PlanetScale Postgres single-node databases start at $5/month on the Base plan, and the PS-5 size provides 1/16 vCPU and 512 MB memory. Single node means one primary and no replicas, so it is explicitly non-HA. PlanetScale says this mode is intended for development, early-stage projects, testing, proof of concepts, and low-traffic production workloads. Single node is available only on network-attached storage clusters; you cannot create a single-node Metal database.
When to choose
Use this when the workload is cost-sensitive, low-traffic, and can tolerate the absence of replica-based high availability. The decisive factor is whether downtime or failover requirements are loose enough that a single primary is acceptable.
Tradeoffs
Lowest entry cost and simplest way to start on PlanetScale Postgres, but no replicas, no automatic HA failover, and less read scaling headroom than multi-node HA clusters.
Cautions
Do not assume you can combine single node with Metal; PlanetScale documents single node as network-attached-storage only. If you later need HA, PlanetScale supports switching from single node to HA by adding two replicas across three availability zones.
PlanetScale Postgres high-availability cluster
As of 2026-04-08, PlanetScale's HA Postgres configuration is the default production shape: 1 primary plus 2 included replicas across 3 availability zones with a 99.99% SLA. Official pricing shows PS-5 HA at $15/month and PS-10 HA at $30/month on network-attached storage, while Metal starts at $50/month for higher-performance local NVMe storage. HA clusters add replica capacity, automatic failover, and the production posture PlanetScale positions for critical workloads. Storage and egress remain separate pricing factors on network-attached-storage clusters.
When to choose
Use this when the workload is customer-facing, needs automatic failover, or cannot tolerate single-node downtime. The decisive factor is HA and production resilience requirements, not just raw compute size.
Tradeoffs
Higher monthly cost than PS-5 single node, but materially better availability, read capacity, and upgrade headroom. Metal improves performance further, but increases spend and is not available in single-node mode.
Cautions
On the Base plan, each production branch provisions its own production database cluster, so branching can multiply costs. If you move from single node to HA or to Metal, PlanetScale documents that it adds replicas and may copy data onto new instances during the transition.
Try with your AI agent
$ npm install -g pocketlantern $ pocketlantern init # Restart Claude Code, Cursor, or your MCP client, then ask: # "PlanetScale PS-5 Single-Node vs Larger Cluster — when does one get cheaper?"