How to spec/size a load balancer for high availability?

March 18, 2014 7.5k views
Hello, I would like offer high availability and have a load balancer control traffic to my application nodes/droplets. I understand that I can use a droplet as a dedicated load balancer. Is there a best practice or rule of thumb when selecting ram and processor size of a load balancer relative to the amount of users or size of the application nodes themselves? For instance, should the specs of the lb be larger or smaller than the application nodes? Any suggestions? Thanks!
2 Answers
A 512MB load balancer droplet running HAProxy or Nginx is usually enough.
In most cases, the app servers are larger than the load balancers but that completely depends on your setup and traffic.

I'd suggest starting off with 512MB load balancer and scaling up as needed.


Like any other application like Exchange etc there is sizing tool for Loadbalancer. Which will help you to size the loadbalancer depending on the size and load of the user.


Have another answer? Share your knowledge.