name: Sui API Rate Limits
description: >
  Rate limits for public Sui API endpoints maintained by Mysten Labs. The
  public endpoints are shared infrastructure and should not be used as the sole
  RPC provider for high-traffic production applications. For higher throughput,
  developers should use community RPC providers or run dedicated full nodes.
rateLimits:
  - name: Public Endpoint Rate Limit
    description: >
      Rate limit applied to the public Mysten Labs-maintained full node
      endpoints (fullnode.<NETWORK>.sui.io:443). This limit was progressively
      reduced in 2024 to encourage developers to use community-provided RPC
      services. The final target rate was reduced from 120 requests/second
      to 10 requests/second, but the documented current limit is 100 requests
      per 30 seconds.
    url: https://docs.sui.io/references/sui-api/rpc-best-practices
    requests: 100
    period: 30 seconds
    scope: Per IP address
    applies_to:
      - gRPC API (fullnode.mainnet.sui.io:443)
      - JSON-RPC API (deprecated)
    networks:
      - mainnet
      - testnet
    note: >
      Mysten Labs announced a graduated reduction from 120 req/s to 10 req/s
      starting July 8, 2024. Consult the current documentation and provider
      announcements for the latest enforced limits.
    forum_announcement: https://forums.sui.io/t/upcoming-public-rpc-rate-limit-changes/45753

  - name: GraphQL Complexity Limits
    description: >
      GraphQL queries are subject to complexity scoring. The Sui GraphQL service
      limits query depth and complexity to prevent resource exhaustion. Consult
      the GraphQL schema reference for current limits.
    url: https://docs.sui.io/references/sui-graphql
    scope: Per query

recommendations:
  - description: >
      Use multiple RPC providers with failover logic for production
      applications to avoid single points of failure.
  - description: >
      Notify providers at least 48 hours in advance for high-traffic events
      such as NFT launches.
  - description: >
      Request capacity expansion from providers at least one week before
      anticipated demand spikes.
  - description: >
      Run a dedicated full node or contract a professional RPC provider
      for high-stakes applications such as marketplaces and exchanges.