Rate Limiting and Complexity Analysis in GraphQL
What you'll learn: How to protect GraphQL APIs from abusive queries by calculating query cost, analyzing depth, and enforcing per-field limits.
The Problem with GraphQL Rate Limiting
Unlike REST endpoints where each URL represents a bounded operation, GraphQL exposes a single endpoint that can execute arbitrarily complex queries. A malicious or careless client could request deeply nested data or fetch thousands of records in one query, overwhelming your backend. Traditional request-count rate limiting (e.g., "100 requests/minute") doesn't work—one GraphQL query can do the work of hundreds.
Query Depth Analysis
Depth limiting restricts how many levels deep a client can nest fields. For example, a depth limit of 5 would reject this query:
user → posts → comments → author → posts → comments → ...
This prevents infinite recursion attacks and bounds the worst-case database joins.
Complexity Score Calculation
Query complexity analysis assigns a cost to each field before execution. Think of it like a points budget: each field costs points, and you reject queries exceeding your threshold.
- Simple scalar fields: 1 point
- List fields: multiply by estimated size (e.g.,
posts(limit: 100)= 100 points) - Expensive computations: higher custom cost
The total complexity is summed before execution. If it exceeds your limit (say, 1000 points), the query is rejected with an error.
Per-Field Cost Limits
You can also assign maximum costs to specific fields. For example, analytics might have a cost cap of 200, preventing clients from requesting heavy aggregations alongside other data.
This approach aligns rate limiting with actual resource consumption rather than arbitrary request counts.
Key Takeaway:
GraphQL rate limiting requires analyzing query structure—calculating depth, summing field complexity scores, and applying cost budgets—to prevent resource exhaustion from single, expensive queries.