Hi, I was looking at different reverse proxies (Traefik, Caddy, NGINX, Apache (mod_proxy), HAProxy) - I wanted to run two instances of a reverse proxy on different devices for failover. I do not find much about running reverse-proxies in a failover configuration, could someone point me at a resource related to these proxies so I can figure out how to run reverse-proxies in an HA configuration?
Thanks
Edit: The answer seems to be using keepalived, because I will be using this only in my homelab and not exposing it to the internet.
Related links if someone is interested:
You probably need something like Fabio or Nginx to serve as a load balancer in front of the two instances you want to proxy.
Just to be clear, I want an HA setup for my reverse proxies. How will running another reverse-proxy (NGINX) in front of my existing reverse-proxies help me here? This will again become a single point of failure which I’m looking to avoid.
Thanks!
One of the easier options would be something like a DNS-based load balancer with an extremely short TTL (<60 seconds). Cloudflare offers an option that is $5/mo to use.
You would need 2 haproxies to do this.
Otherwise, you’re likely looking into something like keepalived with an elastic/additional/failover IP that can swap between instances if a failover situation is detected
edit: something like this would be what I’m talking about for hetzner: https://vitobotta.com/2020/03/20/haproxy-kubernetes-hetzner-cloud/
Essentially you’ll need some sort of API access to be able to swap the floating IP between the ha proxy instances on demand if keepalived detects a fault and needs to swap the floater to another machine.
Thank you. Since this will only be for my LAN, I’ll start reading up on
keepalived. Thanks for mentioning it.Related links if someone is interested:
Thanks!