5 scenarios + mixed traffic + degradation · 50 VUs · ~30 min

GraphQL Bench

Same schema, same queries, fair fight.

Intel(R) Core(TM) i9-9880H CPU @ 2.30GHz · 16 cores
NestJS
Express
0

req/s on health check

avg 6.87msp99 14.94ms
Fastify
0

req/s on health check

avg 4.66msp99 10.24ms
Mercurius
0

req/s on health check

avg 3.04msp99 7ms
Yoga
0

req/s on health check

avg 7.39msp99 16.24ms

Throughput by Scenario

Requests per second — lightest to heaviest

Scenario Deep Dive

Express + Apollo
0req/s

+8% vs slowest

avg 6.87msmed 7.41msp95 9.31msp99 14.94ms
Fastify + Apollo
0req/s

+58% vs slowest

avg 4.66msmed 5.04msp95 6.8msp99 10.24ms
Fastify + Mercurius
0req/s

+140% vs slowest

avg 3.04msmed 3.29msp95 4.89msp99 7ms
Fastify + Yoga
0req/s

baseline

avg 7.39msmed 7.74msp95 11.68msp99 16.24ms

Latency Percentiles — Health Check (Scalar)

Response time distribution — lower is better

Coming Soon

Go

chi · Gin · Fiber

C#

.NET + Hot Chocolate

Java

Spring + Netflix DGS