name: KeystoneJS FinOps
description: >-
  KeystoneJS is MIT-licensed open-source software with no licensing or
  subscription costs. FinOps considerations relate entirely to the
  infrastructure on which Keystone is deployed (compute, database, storage,
  CDN) and any optional commercial support engagement with Thinkmill. Cost
  optimization focuses on right-sizing the hosting environment and using managed
  database services efficiently.
url: https://keystonejs.com
created: "2026-06-13"
modified: "2026-06-13"
softwareCosts:
  licensingCost: Free (MIT open-source license)
  subscriptionCost: None
  enterpriseSupport: Contact Thinkmill for commercial pricing (https://www.thinkmill.com.au/services/keystone)

infrastructureCosts:
  description: >-
    All costs are incurred at the infrastructure level chosen by the operator.
    KeystoneJS runs on Node.js and requires a database (PostgreSQL recommended
    for production). Typical deployment stacks include a Node.js application
    server, a managed PostgreSQL database, and optional object storage for
    images and files.
  components:
    - name: Node.js Application Server
      description: >-
        Keystone runs as a Node.js process. Minimum recommended: 1 vCPU,
        512 MB RAM for small projects. Production workloads typically need
        1-2+ vCPUs and 1-2 GB RAM. Can be deployed to any Node.js-compatible
        host including Railway, Render, Fly.io, AWS EC2/ECS, Google Cloud Run,
        or Heroku.
      estimatedMonthlyCost: "$5–$100+ depending on host and instance size"

    - name: Database (PostgreSQL / MySQL / SQLite)
      description: >-
        PostgreSQL is the recommended production database. SQLite is suitable
        for development and embedded use cases. Managed PostgreSQL services
        include Supabase, Neon, Railway Postgres, AWS RDS, and Google Cloud SQL.
      estimatedMonthlyCost: "$0 (Supabase/Neon free tiers) – $50+ for managed production instances"

    - name: File and Image Storage
      description: >-
        The image and file field types can store assets locally or via
        cloud storage such as AWS S3 or Cloudflare R2 using community
        adapters. Storage costs depend on asset volume.
      estimatedMonthlyCost: "$0–$20+ depending on storage volume and CDN usage"

costOptimizationTips:
  - Use Neon or Supabase free-tier PostgreSQL for development and small projects to eliminate database costs.
  - Deploy the Node.js server to a platform with a free tier (Railway, Render, Fly.io) for low-traffic applications.
  - Enable Prisma query optimization and use the take limit (graphql.queryLimits.maxTotalResults) to prevent expensive runaway queries that inflate compute costs.
  - Use SQLite for local development and embedded read-heavy sites to remove the need for a separate database server.
  - Cache GraphQL responses at the CDN or reverse-proxy layer to reduce compute load per request.
  - Monitor slow queries via database-level tooling (pg_stat_statements on PostgreSQL) rather than relying on application-level instrumentation.

references:
  - url: https://keystonejs.com/docs
    description: Official Keystone documentation including deployment guidance
  - url: https://www.thinkmill.com.au/services/keystone
    description: Thinkmill enterprise support and consulting services
  - url: https://keystonejs.com/blog
    description: Keystone blog with deployment examples and best practices