Hosted server returning localhost/web instead - docker

Question

I have a website host on a server in Digital Ocean that is behaving weirdly.
The website is written in Flask which is deployed in Docker and using reverse proxy with a combination of Let's Encrypt to host on the web.
The website's domain is mes.th3pl4gu3.com.
If I go on mes.th3pl4gu3.com/web/ the website appears and works normal.
If I go on mes.th3pl4gu3.com/web it gives me http://localhost/web/ in the URl instead and conenction fails.
However, when I run it locally, it works fine.
I've checked my nginx logs, when i browse mes.th3pl4gu3.com/web/ the access_logs returns success but when i use mes.th3pl4gu3.com/web nothing comes to the log.
Does anyone have any idea what might be causing this ?
Below are some codes that might help in troubleshooting.
server {
server_name mes.th3pl4gu3.com;
location / {
access_log /var/log/nginx/mes/access_mes.log;
error_log /var/log/nginx/mes/error_mes.log;
proxy_pass http://localhost:9003; # The mes_pool nginx vip
}
listen 443 ssl; # managed by Certbot
ssl_certificate /etc/letsencrypt/..........
ssl_certificate_key /etc/letsencrypt/........
include /etc/letsencrypt/.........
ssl_dhparam /etc/letsencrypt/......
}
server {
if ($host = mes.th3pl4gu3.com) {
return 301 https://$host$request_uri;
} # managed by Certbot
listen 80;
server_name mes.th3pl4gu3.com;
return 404; # managed by Certbot
}
Docker Instances:
7121492ad994 docker.pkg.github.com/mervin16/mauritius-emergency-services-api/mes:1.2.5 "uwsgi
app.ini" 4 weeks ago Up 4 weeks 0.0.0.0:9002->5000/tcp mes-instace-2
f4dc063e33b8 docker.pkg.github.com/mervin16/mauritius-emergency-services-api/mes:1.2.5 "uwsgi app.ini" 4 weeks ago Up 4 weeks 0.0.0.0:9001->5000/tcp mes-instace-1
fb269ed2229a nginx "/docker-entrypoint.…" 4 weeks ago Up 4 weeks 0.0.0.0:9003->80/tcp nginx_mes
2ad5afe0afd1 docker.pkg.github.com/mervin16/mauritius-emergency-services-api/mes:1.2.5 "uwsgi app.ini" 4 weeks ago Up 4 weeks 0.0.0.0:9000->5000/tcp mes-backup
docker-compose-instance.yml
version: "3.8"
# Contains all Production instances
# Should always stay up
# In case both instances fails, backup instance will takeover
services:
mes-instace-1:
container_name: mes-instace-1
image: "docker.pkg.github.com/mervin16/mauritius-emergency-services-api/mes:${MES_VERSION}"
networks:
- mes_net
volumes:
- ./data:/app/data
env_file:
- secret.env
ports:
- "9001:5000"
restart: always
environment:
- MES_VERSION=${MES_VERSION}
mes-instace-2:
container_name: mes-instace-2
image: "docker.pkg.github.com/mervin16/mauritius-emergency-services-api/mes:${MES_VERSION}"
networks:
- mes_net
volumes:
- ./data:/app/data
env_file:
- secret.env
ports:
- "9002:5000"
restart: always
environment:
- MES_VERSION=${MES_VERSION}
networks:
mes_net:
name: mes_network
driver: bridge
docker-compose.yml
version: "3.8"
# Contains the backup instance and the nginx server
# This should ALWAYS stay up
services:
mes-backup:
container_name: mes-backup
image: "docker.pkg.github.com/mervin16/mauritius-emergency-services-api/mes:${MES_VERSION}"
networks:
- mes_net
volumes:
- ./data:/app/data
env_file:
- secret.env
ports:
- "9000:5000"
restart: always
environment:
- MES_VERSION=${MES_VERSION}
nginx_mes:
image: nginx
container_name: nginx_mes
ports:
- "9003:80"
networks:
- mes_net
volumes:
- ./nginx/nginx.conf:/etc/nginx/conf.d/default.conf
- ./log/nginx:/var/log/nginx
depends_on:
- mes-backup
restart: always
networks:
mes_net:
name: mes_network
driver: bridge
I have multiple instances for load balancing across apps.
Can someone please help or if anyone has any clue why this might be happening ?

Answer 1

As long as I tested the page https://mes.th3pl4gu3.com/web with or without / trailing at the end it worked fine. (Firefox version 87 on Ubuntu)
Maybe there is a bug / problem with your web browser or any kind of VPN / Proxy you are running. Make sure all of them are off.
Plus on Nginx you can get rid of trailing / using rewrite rule
e.g.
location = /stream {
rewrite ^/stream /stream/;
}
which tells Nginx parse stream as it is stream/
and for making sure you are not facing any issue because of cached data, disable and clear all the cache. On your web browser hit F12 -> go to console tab , hit F1 and there disable cache. On Nginx set "no cache" for header, e.g.
add_header Last-Modified $date_gmt;
add_header Cache-Control 'no-store, no-cache, must-revalidate, proxy-revalidate, max-age=0';
if_modified_since off;
expires off;
etag off;

Answer 2

I tested your site with Chrome, Safari, and Curl, and I can't see that issue.
Try clear your cache.
Method 1: Ctrl-Shift-R
Method 2: DevTool -> Application/Storage -> Clear site data

Answer 3

My guess is it is related to your Flask SERVER_NAME
As you said, locally your SERVER_NAME might be set to localhost:8000.
However, on production, it would need to be something like
SERVER_NAME = "th3pl4gu3.com"
Your issue is that you are pulling the SERVER_NAME from the Flask SERVER_NAME variable, so it ends up as https://localhost/web instead of your desired URL.