keepalive

**Jetty version(s)**
9.4.29 10.0.26

**Jetty Environment**

**HTTP version**
HTTP 1.1

**Java version/vendor** `(use: java -version)`
java11

**OS type/version**
AWS EC2

**Description**
_((Google Translated))_
nginx keeaplive 65s

jetty idletimeout configured 70s

When data packets arrive at the network card, Jetty occasionally takes several hundred milliseconds to process them. Our service has a high request volume, occurring several times per minute. We have an 8-core CPU.

After changing selectors to 8, the frequency of this occurrence decreased.

Without changing selectors,

nginx keeaplive 10s

jetty idletimeout configured 16s

The idle time is almost gone.

**How to reproduce?**

Start by reviewing Jetty 9.4.29 and 10.0.26 selector handling together with the configured idle-timeout paths, then compare the reported nginx keepalive and Jetty timeout settings. Reproduce the behavior with HTTP/1.1 traffic on AWS EC2; done requires identifying why packet processing is delayed and validating a corrective configuration or code change.

Written by the indexing model from the issue text.