How to set up two Nginx containers as a reverse proxy in an active-passive set up with failover? - nginx

Question

I have set up a Nginx container on a Linux-EC2 server. My Nginx config file is as follows:
server {
listen 80;
server_name client-dev.com;
location / {
proxy_pass http://dev-client.1234.io:5001/;
proxy_redirect off;
##proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
client_max_body_size 10m;
client_body_buffer_size 128k;
proxy_connect_timeout 90;
proxy_send_timeout 90;
proxy_read_timeout 90;
proxy_buffer_size 4k;
proxy_buffers 4 32k;
proxy_busy_buffers_size 64k;
proxy_temp_file_write_size 64k;
}
}
server {
listen 80;
server_name client-test.com;
location / {
proxy_pass http://test-client-1234.io:5005/;
proxy_redirect off;
##proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
client_max_body_size 10m;
client_body_buffer_size 128k;
proxy_connect_timeout 90;
proxy_send_timeout 90;
proxy_read_timeout 90;
proxy_buffer_size 4k;
proxy_buffers 4 32k;
proxy_busy_buffers_size 64k;
proxy_temp_file_write_size 64k;
}
}
It passes the requests with different ports to port 80. Now, I need to make a redundant Nginx in an active-passive mode. In case, if the Nginx container goes down/stops.
To do so, would I need to set another Nginx container in the same server? If so, how should it be setup to do the failover automatically?
I have looked at "upstream" option, but as I found, it would not work for this case. The proxy_pass that I have are external and dynamic that I get them using a script from docker-cloud.
There is another way named "docker-gen" however, I'm not sure how much useful it would be, and I prefer to use another way if there is any?
Any help would be appreciated.

Answer 1

I can think of following options:
Kubernetes: You can create a deployment for your nginx setup and use liveliness probes. Kubernetes will probe the nginx container with http request/interval you provide, if the pod is not healthy, it will be killed and recreated. Using multi nodes in your kubernetes cluster you can even mitigate from node failure.
Docker Swarm: Using docker swarm mode with multiple nodes, you can mitigate from node failure, but nginx health should be checked by an external custom script that can be done with bash and curl.
Using standalone hosts with keepalived: This is traditional nginx active/passive cluster using keepalived. You can also use this with docker but it would be dirty. because all of your containers will be passive on one host.