How We Calculate Serverless Database Pricing Estimates

Our goal is to give you a consistent, reproducible way to compare how pricing scales across different serverless database platforms.

For the actual pricing comparisons page, see Reference → Serverless Database Price Comparison.

Standardized Assumptions

The workload model

To keep comparisons consistent, we use a unified workload model (exceptions only where enforced):

• Free-tier allowances are deducted before calculating paid usage.
• Storage is measured as total provisioned dataset size, including indexes and overhead.
• Compute is measured as monthly query volume unless the provider charges by compute unit (vCPU/RAM hours).
• Data transfer costs and egress are excluded from per-provider calculations.

The usage tiers

These tiers mirror our Functions Pricing page structure so comparisons stay consistent across content types:

• Tier 1: Free tier break-even
  The moment your usage exceeds the monthly free allowance for storage, compute, or row limits. Equivalent to the functions pricing Free Tier break-even.

• Tier 2: 10 GB database
  A moderate workload with 10 GB storage. Exposes per-unit pricing, tier lock-in, and storage cost differences. Equivalent to the functions pricing 50M invocations tier.

• Tier 3: 100 GB database
  A heavy sustained workload where HA, replicas, storage costs, and data transfer dominate. Equivalent to the functions pricing 500M invocations tier.

These tiers are intentionally chosen to reveal scaling behavior, not to judge which provider is “cheapest.”

What inputs are used

Each platform is calculated using the values explicitly published in its pricing documentation:

• Storage cost per GB per month (included vs overage)
• Compute cost per hour or per million queries
• Free tier allowances (storage, compute hours, row limits)
• Minimum commitment or baseline spend
• Replica and high availability costs
• Backup storage costs
• Connection limits and their upgrade triggers

What this methodology does not cover

These estimates do not include:

• Egress and data transfer fees
• Cold-start latency or connection pool limits
• Regional pricing differences
• Vendor lock-in or migration costs
• Integrated services (search, auth, real-time subscriptions) unless bundled
• Discounts, reserved capacity, and enterprise pricing are excluded
• Managed services that also provide compute (e.g. Railway, Fly.io Postgres) are not included as pure DBaaS

Why these numbers can differ from your bill

Database pricing depends strongly on workload patterns. Real databases vary in:

• Read/write ratio (writes are typically more expensive)
• Index sizes and storage amplification
• Connection pool utilization
• Backup retention policies
• Replication factor and multi-region setup

This methodology removes that variability so costs stay comparable across providers. Always validate against your actual usage patterns.

For the computed pricing table, see Reference → Serverless Database Price Comparison.

Source data

All pricing data is sourced from official provider pricing pages and fetched using Lightpanda (headless Chromium) and web_fetch. If a provider updates its pricing, the reference table is updated accordingly.