I had the same issue. I decided to install nginx on each droplet and have that handle proxying from port 80 to my node app’s port (4000, in my case). I noticed that after installing - but before configuring the proxy - the load balancer was showing the default nginx welcome page.

This indicates to me that the load balancer 80->4000 forwarding option is not taking any effect at all, and I was previously seeing the 503 error because it was hitting my droplets on port 80.

I’m going to leave the 80->4000 setting in place for now, even after setting up nginx. It’s possible that this forwarding option just takes some time to kick in? I’ll report back if I find out how to do this without nginx. (We’ll see what happens when I get into SSL, too)

What is the configuration in the load balancer?
Did you configure a port for SSL on the Droplet (not the load balancer)?

I did’t configure either of the droplets or the load balancer for SSL. Doing everything over plain HTTP and the droplets worked fine when I reach them via HTTP by themselves. I’ve actually switched to creating my own loading balancing droplet because I couldn’t get this to work. I setup a vanilla droplet with Nginx as revers proxy load balancing my droplets. That’s working fine for me.

I am having this same issue.... My Load Balancer shows healthy, but it’s like it doesn’t want to pass “real” traffic (outside of the health check) to the droplet… Has there been a fix for this?