haproxy

It just means that instead of deploying Baker Street as part of your microservice push, you're deploying a) your microservice b) your dual HAProxy setup and/or new HAProxy config and c) DNS.

I am acutely aware of HAProxy however the rest of the team is, no surprise I'm sure, familiar with nginx.

I wrote a simple Python script (92 lines of code) that spits out a new haproxy configuration and gracefully reloads haproxy whenever the DNS changes.

We are using CoreOS for the host servers, dumping logs to logstash, and connect the containers via HAProxy.

How does Baker Street handle restarting haproxy, does it do anything like this[0] automatically to get zero-downtime configuration reloads?

Bonus: haproxy is a superior load balancer with better runtime metrics and health checking options than nginx.

Deploying new microservices is as simple as managing the HAProxy config file via Chef or Puppet.

Curious: why choose something like Watson over using haproxy's pretty solid built in health checking mechanisms?

Except most people deploy dual HAProxy servers connected via heartbeat, which share a floating IP.

Nginx still has a lot of work to do to match haproxy in this area.

What I settled on was the following: Ngninx for SSL termination in front of an haproxy load balancer.

There is more programming involved, since you need to figure out how to tie a new instance deployment into Chef/Puppet/etc to update HAProxy.

This is more of a frustration with nginx which could be so much more for us if: * They were not purposely holding back features so they can be pay-walled * Re-compiling were not necessary to add functionality * More cohesive modification ecosystem I know I said "reverse proxy" which HAProxy is cleary the more superior, but I didn't mean to limit my statement to reverse proxies.