Flags SDK Improves Feature Flag Evaluation Speed by 10x

Flags SDK and Vercel Flags have improved the speed of bulk-evaluating multiple feature flags by approximately 10x compared to before. This performance improvement was achieved by reducing microtask queue overhead and minimizing the number of Promises generated, and the greater the number of flags to be evaluated, the larger the performance gain.

To take advantage of the optimized performance, instead of the existing Promise.all([flagA(), flagB()]) approach, you should use await evaluate([flagA, flagB]) with array destructuring or await evaluate({ a: flagA, b: flagB }) with object destructuring.

Also, this performance improvement is automatically applied to the precompute() function as well. You can start using the improved performance right away by updating to the latest versions of the flags and @flags-sdk/vercel packages.