Nginx creates more connection than poolsize, backlog to redis

I am using https://github.com/openresty/lua-resty-redis#connect method and keepalive()
i have kept pool size to 200 and backlog 20, Still it creates 2000 connections to redis during a load test.
is_connected, err = client:connect(REDIS_SERVER1,REDIS_PORT,{ pool_size = 200, backlog = 10})

ngx-lua : 5.1
workers : 2

What could be the reason for more number of connection from nginx to redis ? How can we control or restrict it ?

Used Nginx conf

user              www-data;
worker_processes  2;
worker_rlimit_nofile  10240;

include /etc/nginx/modules-enabled/*.conf;

error_log         /var/log/nginx/error.log;
pid               /var/run/nginx.pid;

events {
  worker_connections  10240;
  multi_accept  on;
  use epoll;
 }
 http {
  
    server_tokens       off;

    include       /etc/nginx/mime.types;
    default_type  application/octet-stream;

    root        /srv/www/default;

    sendfile    on;
    tcp_nopush  on;
    tcp_nodelay on;

    keepalive_timeout   120;
    keepalive_requests  1000;
    reset_timedout_connection on;

    server_names_hash_bucket_size 128;
    types_hash_max_size 2048;
    types_hash_bucket_size  64;

    open_file_cache max=200000 inactive=20s;
    open_file_cache_valid 30s;
    open_file_cache_min_uses 2;
    open_file_cache_errors on;

    client_body_timeout   5;
    client_header_timeout 5;
    client_header_buffer_size 256;
    client_max_body_size  16m;

    ##
    # Gzip Settings
    ##

    gzip            on;
    gzip_http_version 1.0;
    gzip_buffers    4 16k;
    gzip_min_length 1024;
    gzip_comp_level 2;
    gzip_proxied    expired no-cache no-store private auth;
    gzip_types      text/plain text/css text/javascript application/javascript application/x-javascript text/xml application/xml application/xml+rss application/json;
    gzip_disable    "MSIE [1-6]\.(?!.*SV1)";

    set_real_ip_from    10.8.0.0/16;
    real_ip_header      X-Forwarded-For;
    real_ip_recursive   on;

    include /etc/nginx/conf.d/*.conf;

    

    # proxy settings
    proxy_buffer_size       16k;
    proxy_buffers           8   16k;
    proxy_connect_timeout   60s;
    proxy_send_timeout      60s;
    proxy_read_timeout      60s;

    # local log
    access_log  /var/log/nginx/$host.access.log ltsv;

    ##
    # Virtual Host Configs
    ##

    include /etc/nginx/sites-enabled/*;
 }

Steps To Reproduce:

1. Take a 2 core CPU machine
2. Use resty.redis lua
3. configure poolsize 200 and backlog 10
4. Do a stress test to generate 1000 req/s to nginx.
5. check the connected clients in redis to verify
6. check netstat | grep '6379' to check open connections.

Start with the lua-resty-redis connect() and keepalive() documentation and the reported pool_size/backlog configuration. Reproduce the 1,000 requests-per-second load described, then compare Redis connected clients with netstat on the two-worker Nginx setup. Done means identifying whether the observed connections match documented pooling behavior and documenting or fixing a reproducible cause.