Installing Discourse and Ghost on the same Droplet

AI-generated summary

Users discuss the feasibility of installing Discourse and Ghost on the same Droplet. Alaa asks if it’s safe, and itsbhanusharma suggests it’s possible but not ideal. pfaffman confirms it’s doable, recommending Traefik as a reverse proxy and noting a minimum of 2GB RAM is needed, though two smaller droplets may be easier and more cost-effective. justin agrees, emphasizing that reverse proxy setup justifies the extra cost. ariznaf shares experience running both on a 1GB, 1-core Lightsail server, noting high CPU usage and performance issues, and suggests 2 cores and 2GB RAM are more appropriate. They also mention that while a t2.micro works well, AWS pricing complexity makes cost prediction difficult, leading to a consideration of separate instances versus upgrading to a more powerful server.

Can I safely install Discourse and Ghost on the same Droplet?

You probably can (if you know how to) but it won’t be ideal.

Certainly. I’ve been very happy with traefik as a front end /reverse proxy and configuring it with docker environment variables. You’ll need at least 2gb of ram, so it’s not going to save you money at the low end of the spectrum. For low volume, 2 small droplets is much easier.

Agreed. The reverse proxy setup makes the slightly higher monthly cost for two droplets worth it by itself. I always get bit by those…

I have done that in an amazon lightsail server.

I have installed discourse (running under docker of course) then I have installed the docker version of ghost using docker-compose and creating a stack.yml file to configure it.

Then I had installed nginx and used it as a front end, configuring discourse serving content in a unix pipe and redirecting calls to the ghost server to the internal 0.0.0.0:2386 address where ghost listens.

I am looking now to see if I can serve ghost content in a unix pipe too.

I have a 1 Gb and 1. Vcore server.
But it does not run ok. It hoes above the sustain cpu usage limit.
At least ehen you need to reconfigure or rebuild, and it gets way too slow.
It seens a problem of cpu usage, nor ram limit.
2 cores and 2 gb ram seems more appropiate.

In a t2.micro with same config it seems to run smoothly.
But aws pricing is way too complex to be able to predict costs.

Ee have now to evaluate if running separate instances of ghost and discourse us cheaper and more appropuate that going for a 2core 4gb memory lightsail server.