bug: Performance on container is no good, just 5.6k QPS

[Mararesliu]

Current Behavior

I run apisix on one pod with two container, limit 2 cpu, benchmark from apisix performance just 5.6k QPS, but benchmark to upstream directly is ok(26k). The problem is upstream latency too high, I see on the dashbord metrics

I use Envoy ,it can got more than 8k QPS.

Expected Behavior

How can in increase upstream performance via apisix?

Error Logs

No response

Steps to Reproduce

1、RUN a pod with apisix and nginx demo.
2、setting root route to nginx
3、benchmark to apisix like
./wrk -t600 -c5000 -d120s -H "Host: nginx" http://172.17.0.2:9080

Environment

• APISIX version (run apisix version): 3.11
• Operating system (run uname -a): redhat 8.11
• OpenResty / Nginx version (run openresty -V or nginx -V):
• etcd version, if relevant (run curl http://127.0.0.1:9090/v1/server_info):
• APISIX Dashboard version, if relevant:
• Plugin runner version, for issues related to plugin runners:
• LuaRocks version, for installation issues (run luarocks --version):

[moonming]

For API Gateway performance testing, please ensure that the upstream is not a performance bottleneck, otherwise the performance cannot be fully released. The ideal test scenario is that the CPU of the API Gateway data plane is 100% used, refer to https://docs.api7.ai/apisix/production/performance/performance-testing

[Mararesliu]

It's because of i enable jager plugin.