Heroku Performance-M pricing & rate limits

Heroku Postgres occupies a premium price bracket compared to bare-metal VPS providers like Linode or Vultr, focusing on managed high-availability rather than raw hardware cost-efficiency. Entry-level production tiers start at $250.00/mo, which provides a modest 2.5 GB of RAM and 250 GB of disk storage backed by 1 vCPU. Scaling up to the high-end $1500.00/mo tier grants access to 126 GB of RAM, 2000 GB of storage, and 16 vCPU cores. This architecture is designed for teams that prioritize managed replication and automated failover over the granular control offered by self-hosted Postgres on AWS EC2 or EPYC-based dedicated servers.

• Database connection limits start at 120 concurrent connections for the base tier, scaling up to 500 on top-tier instances.
• Storage expansion is not handled via standard block storage attachments; users must scale the entire database tier to increase capacity, which avoids the complexity of manual volume management but inflates monthly costs.
• High-availability is baked into the platform architecture, utilizing follower nodes for read-scaling and automated failover, which mitigates the risk of downtime inherent in single-node database deployments.
• Excessive storage usage beyond the 2000 GB limit triggers a move to private, custom-quoted plans, effectively ending the predictable self-service pricing model.

The platform is suited for applications where database maintenance is an unwanted distraction for the engineering team. Teams relying on complex extensions or requiring deep kernel-level tuning will find the abstraction restrictive, but those needing a reliable, "set-and-forget" PostgreSQL environment will find the cost justified by the elimination of manual vacuuming, backup scheduling, and replication monitoring. It remains a stable choice for scaling web backends that have outgrown simple VPS setups but are not yet ready to manage a full-scale RDS or self-hosted Kubernetes database cluster.