GuzzleHttp cause 504 and upstream timeout from one droplet behind the load balancer

I got suddenly my server can not be access…right now just for test i have one droplet behind the load balancer and firewall with standard setup (Digital Ocean FW to that droplet), when i see on nginx/log/access i got the access that same with the error time is like this :

[09/Mar/2022:08:20:53 +0000] “GET /ap?co=42656._px&cid=342888 HTTP/1.1” 504 176 “-” “GuzzleHttp/7”

and the error log is

2022/03/09 08:20:53 [error] 2323#2323: *1232 upstream timed out (110: Connection timed out) while reading response header from upstream, client: 11.232.0.1, server: domain.com, request: “GET /ap?co=42656._px&cid=342888 HTTP/1.1”, upstream: “fastcgi://unix:/var/run/php/php7.4-fpm.sock”,

so i think it’s something wrong on load balancer or from firewall? is using guzzlehttp / curl to outside domain to get data considering as “outbound” ?

so i am duplicate this server and put it as stand alone droplet and pointing the domain to there…when it access like that everything run ok? no error at all and run smooth even i leave for one day…this is duplicate server from snapshot so nginx config,etc should be same…the different just not behind the loadbalancer? any idea why? thank you!