Has anyone encountered this problem? I'm experiencing a brief 502 error

Projects NGINX
, ,
1

Please use this template for troubleshooting questions.

**My issue:**Has anyone encountered this problem? I’m experiencing a brief nginx 502 error, but my backend is working fine.

**How I encountered the problem:**Last night, my nginx encountered a 502 error, indicating no healthy upstream and preventing requests from being sent to any backend. This issue only lasted about 7 seconds. My nginx is behind a GCP load balancer.

Solutions I’ve tried:I did nothing.

**Version of NGINX or NGINX adjacent software (e.g. NGINX Gateway Fabric):**nginx/1.26.2

**Deployment environment:gcloud vm instance,**Ubuntu 24.04.1 LTS

Minimal reproducible NGINX config:

worker_processes auto;

events {
worker_connections 1024;
}

http {

limit zones for request rate

limit_req_zone $server_name zone=global_limit:10m rate=200r/s;
limit_req_zone $request_uri zone=uri_limit:10m rate=50r/s;

upstream pay-api-open {
    server 10.149.0.21:18092;
    server 10.149.0.22:18092;
    keepalive 65;
}

server {
    listen 80;
    server_name example.id prod.aaa.id;

    # Paths where issue occurs
    location ~ ^/(gateway/order/detail/v2|api/digitCoin/coinList|gateway/getPayMethod/v2|gateway/order/status|gateway/preserveApp) {
        limit_req zone=global_limit burst=2000 nodelay;
        limit_req zone=uri_limit burst=500 nodelay;
        limit_req_status 430;

        proxy_pass http://pay-api-open;
        proxy_http_version 1.1;
        proxy_set_header Connection "";
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
    }

    location / {
        proxy_pass http://pay-api-open;
        proxy_http_version 1.1;
        proxy_set_header Connection "";
        proxy_set_header Host $host;
    }
}

}

NGINX access/error log: (Tip → You can usually find the logs in the /var/log/nginx directory.)
access.log:“2025-10-30 22:49:49” - “8.215.82.129” - “prod.aaa.id” - “python-requests/2.32.3” - “ID” - “” - “0.000” - “pay-api-open” - “0.000” - “POST /gateway/prepaidOrder HTTP/1.1” - “502” - “150” - “” - “” - “172.17.1.13”
error.log: 2025/10/30 22:49:55 [error] 2184104#2184104: *32398187 no live upstreams while connecting to upstream, client: 172.17.1.10, server: example.id, request: “POST /gateway/getPayMethod/v2 HTTP/1.1”, upstream: “https:pay-api-open/gateway/getPayMethod/v2”, host: “test.aaa.id”

2

Hey @jufu_li!

By all accounts, my initial suggestion would be that the backend stopped being healthy for the 7 seconds the requests failed. Is there any chance that could have been the case?

3

It’s unlikely that both backends would become unhealthy within 7 seconds. The problem is more likely that a backend API is experiencing continuous timeouts, causing Nginx to conclude that both backends are faulty.

4

NGINX determines the health of each upstream server separately, so that is unlikely to have been the case :slight_smile:

5

Thank you for your reply. We are trying to reproduce the result.

1 Like