Use NGINX as database proxy

#sre #devops #kubernetes

In our everyday life as dev, we can lose temporary an access to a database from our workstation and it's still available from our Kubernetes cluster. So, in this kind of situation, we can use a proxy to access the database to resolve this network issue.

So today, we will see how to do it with Nginx.

What is NGINX ?

For the ones who didn't already know what is Nginx, it's an opensource software which can be used as web server, load balancer or proxy.

Configurate NGINX

For our case, the Nginx configuration should look like this.

user  nginx;
worker_processes  auto;
error_log /dev/stdout info;
pid        /var/run/nginx.pid;
events {
  worker_connections  1024;  ## Default: 1024
}
http {
  access_log /dev/stdout combined;
  log_format logger-json escape=json '{"source": "nginx", "time": $msec, "resp_body_size": $body_bytes_sent, "host": "$http_host", "address": "$remote_addr", "request_length": $request_length, "method": "$request_method", "uri": "$request_uri", "status": $status,  "user_agent": "$http_user_agent", "resp_time": $request_time, "upstream_addr": "$upstream_addr"}';
    server {
         listen 8080 so_keepalive=on;
        location = /health {
          auth_basic off;
          return 200;
        }
    }
}
stream {
    upstream database {
        server [database endpoint]:[database port];
    }
    server {
        listen [database port] so_keepalive=on;
            proxy_connect_timeout 60s;
            proxy_socket_keepalive on;
            proxy_pass database;
    }
}

Before to use it, some elements must be replaced:

[database endpoint] : The url to access the database
[database port] : The port to access the database

Detail of the configuration

http

This part is only to expose a /health endpoint to use it for the readiness and liveness probes.

stream

This part is the major one in our case.

upstream is here to create a server group. (We don't have multiple instances here, but it makes easier the configuration to read)

The server part will define :

on which port Nginx must listen -> listen
where the traffic will be redirected -> proxy_pass And other configurations like so_keepalive=on and proxy_socket_keepalive on; which are really importante to keep the connection alive.

Deploy on Kubernetes

Once the Nginx configuration is ready, we can now deploy a Nginx instance in our cluster.

Here is a template to create the configmap, deployment & service.

apiVersion: v1
data:
  nginx.conf: "user  nginx;\r\nworker_processes  auto;\r\nerror_log /dev/stdout info;\r\npid
    \ /var/run/nginx.pid;\r\nevents {\r\n  worker_connections  1024;  ## Default:
    1024\r\n}\r\n\r\nhttp {\r\n  access_log /dev/stdout combined;\r\n  log_format
    logger-json escape=json '{\"source\": \"nginx\", \"time\": $msec, \"resp_body_size\":
    $body_bytes_sent, \"host\": \"$http_host\", \"address\": \"$remote_addr\", \"request_length\":
    $request_length, \"method\": \"$request_method\", \"uri\": \"$request_uri\", \"status\":
    $status,  \"user_agent\": \"$http_user_agent\", \"resp_time\": $request_time,
    \"upstream_addr\": \"$upstream_addr\"}';\r\n\r\n\r\n    server {\r\n         listen
    8080 so_keepalive=on;\r\n\r\n        location = /health {\r\n          auth_basic
    off;\r\n          return 200;\r\n        }\r\n    }\r\n}\r\nstream {\r\n    upstream
    postgres {\r\n        server database:port;\r\n