Why Dokploy Uses So Much Memory

[sonmezerekrem]

Hello everyone,

I have a situation with latest releases of dokploy starting around v0.27-28. I have deployed multiple instances of dokploy for different clients. I am using VPS machines from Hetzner. I have tried 4gb and 16gb vps instances. The issue is about the memory. I am seeing over 1 GB memory usage in most instances, if not minimum 800mb. I think its too much for a management system.

What is your experience with memory usage of Dokploy?

[agentHits]

I have a lot of projects, it's used well, I'm still improving the project in my separate fork, since the admin takes a long time to respond and react to my suggestions or improvements, as well as to my PR - the admin is too slow)

[agentHits]

@sonmezerekrem @Siumauricio

[imrja8]

If you are looking at the Dokploy container specifically, it typically only takes around 400MB - 800MB. (Keep in mind that Linux and Docker will naturally allocate more memory for caching if your server has more RAM available, this is a good thing and improves performance).

However, whatever numbers you are looking at might be the total memory usage of the server after the complete Dokploy installation.

Just for some context on how Linux servers handle memory:

• A fresh, blank Ubuntu server uses about 200MB - 300MB at a normal idle state.
• Docker is naturally memory hungry. Installing the Docker daemon engine alone pushes the server memory usage up to ~500MB before you even run a single container.
• To ensure maximum stability, performance, and internal routing, a full Dokploy installation doesn't just run the Dokploy app. It automatically spins up Traefik (for routing), PostgreSQL (for the internal database), and Redis (for queuing and caching).

When you combine the base OS, the memory-hungry Docker engine, Traefik, Postgres, Redis, and the Dokploy container itself, seeing higher memory usage on your server dashboard is completely normal and expected!

So, seeing ~1GB of RAM usage on a fresh 4GB server with a clean Dokploy installation is completely normal. For comparison, I personally use an AWS c8a.4xlarge (32GB RAM) server, and because Linux and Docker aggressively cache when more RAM is available, my server used 1.7GB to 1.8GB of RAM with a clean Dokploy installation.