Why would Apache Benchmark say all requests completed successfully despite ListenDrops increasing?

Question

I am stress testing a flask application using ab -n 10000 -c 1000 http://192.168.1.16:9090/, and monitoring ListenDrops with nstat -az TcpExtListenDrops on Ubuntu 22.04 (Kernel: 6.8.0-49-generic). My understanding is, each ListenDrops indicates a request that should be discarded, and the client should receive either an error, or no response at all. Despite this, ab says everything was a success.

Concurrency Level:      1000
Time taken for tests:   28.579 seconds
Complete requests:      10000
Failed requests:        0

TcpExtListenDrops increases by nearly 10,000, so what is going on?