After I do pm2 reload, I cannot access my website for 10-15s

Question

I'm currently using cluster mode pm2 and nginx to serve the website. As soon as I deployed a new version with pm2 reload, the website let's say example.com throws 502 bad gateway error. After 10-15s, things become normal and I can access my website.

The error is here

2024/03/25 14:19:48 [error] 22594#22594: *5447 no live upstreams while connecting to upstream, client: 172.71.131.128, server: servers.example.com, request: "OPTIONS /api/serverApi/getServers HTTP/1.1", upstream: "http://servers/api/serverApi/getServers", host: "servers.example.com"

What could be the root problem? I'm stuck here because it functions well after 10-15s. Below is one of my nginx conf file.

upstream websocket {
    #least_conn;
    #ip_hash;
    server 127.0.0.1:1442;
    server 127.0.0.1:1443;
    # server 127.0.0.1:1444;
    # server 127.0.0.1:1445;
    keepalive 64;
}

server {
    listen 80;
    listen [::]:80;
    server_name socket.example.com;

    location / {
        proxy_pass http://websocket;
        include /etc/nginx/proxy_params;
        proxy_connect_timeout 1m;
        proxy_read_timeout 1m;
    }

    location ^~ /nginx_status {
        auth_basic "Admin Login";
        auth_basic_user_file /etc/nginx/pma_pass;

        stub_status on;
        access_log off;
        allow example_ip/32;
        # deny all;
    }

}

I tried adding proxy-timeout but it didn't solve the problem. What can I do more? I want a seamless update instead of 502. It works seamlessly if I go around the website. But if I hard reload F5 the website, it shows 502 for 10-15s.