SingleFlight: The Small Pattern That Quietly Saves Your Application Performance

In modern application development — whether backend APIs or frontend data fetching — you sometimes face many concurrent requests for the same resource. Without careful handling, this can lead to:

• Excessive load on databases or external services
• Duplicate function executions
• Rate limit or performance issues

A smart solution to this is the SingleFlight pattern — not often talked about — but very valuable.

What Is SingleFlight?

SingleFlight ensures that when an asynchronous operation is already in-flight (running) for a given key, additional concurrent calls with the same key don’t trigger a new execution. Instead, they wait for the first execution and share its result.

If that line is invoked 10 times concurrently with "user-123", only one real API/db call runs — and all callers get the same result.

Flyonce Package Overview

flyonce is a TypeScript library develop by wahyu agus arifin, that implements SingleFlight with optional caching. It works with Bun, Node.js, Deno, and other JavaScript runtimes.

Key Features:

• SingleFlight deduplication per key
• Optional result caching with TTL
• Graceful error handling
• Zero dependencies, lightweight
• High performance and concurrency-aware

Contoh

Only one actual API call will run for the key "user-123".

Real-World Use Cases

SingleFlight (and flyonce) is especially helpful for:

• API rate limiting & de-duplication
• Heavy or expensive database queries
• Microservice call optimization
• Cache stampede protection
• Concurrent data fetching in frontends

Without it, concurrent calls will execute independently — wasting CPU, memory, and bandwidth.

Benefits

• Reduces redundant work
• Improves performance
• Ensures consistency
• Easy to integrate

Conclusion

The SingleFlight pattern is a clever way to deduplicate async operations. The flyonce library makes it practical with caching and a clean API. Whether you’re building server systems or frontend apps, flyonce can help keep performance tight and resource usage low.