17 - Nov - 2026

I stopped SSH-ing into Docker just to change my containers

I’ve opened PowerShell, SSH’d into a Linux box, changed one line in a Compose file, and wondered why such a tiny job needed this much ceremony more times than I can count. The terminal is still where I want to be when a server misbehaves, but changing a port or restarting a healthy container hardly qualifies as an emergency.

Dockge takes those routine Docker Compose jobs and puts them in a browser. After running it against a Debian Docker host for an evening, I found myself reaching for SSH considerably less often.

Dockge took about five minutes to make sense

It stays focused on Compose stacks

dockge-compose-file-default-configuration

Dockge is built around Docker Compose stacks, and that’s the first thing you need to know before you install it. It doesn’t try to expose every network, volume, image, and standalone container on the host. Its official documentation is pretty clear about that focus.

That was exactly what I wanted after spending time with browser tools that replaced several of my routine SSH jobs. Portainer is great when I want Docker laid out in exhaustive detail. Dockge has far less furniture.

Installation is short. On a supported Linux host, you create directories for Dockge and its stacks, download the official Compose file, and bring it up with Docker Compose. The default stacks directory is /opt/stacks, while the web interface listens on port 5001.

dockge-container-running-on-port-5001

If you’re working with a fresh server, use Docker’s own documentation to install Docker Engine, then follow Dockge’s setup instructions for the dashboard itself.

dockge-create-admin-account

Dockge’s default Compose configuration mounts /var/run/docker.sock, which gives the application powerful control over the Docker daemon and, by extension, the host. Keep the dashboard on a trusted network or access it remotely through a VPN or a properly secured reverse proxy, rather than exposing port 5001 directly to the internet.

Changing one Miniflux port sold me on it

The boring jobs no longer demand a terminal

For testing, I gave Dockge three stacks instead of tossing it a token hello-world container. Miniflux ran alongside PostgreSQL, Gitea had its own PostgreSQL database, and Uptime Kuma gave me a simpler single-container service.

Dockge found them after I placed each project under /opt/stacks and used Scan Stacks Folder. Existing projects need to follow Dockge’s directory layout, with each compose.yaml living inside its own stack folder. Its migration instructions also recommend stopping a stack before moving the Compose files into that structure and scanning for it.

If you’re putting a server together from scratch, we’ve already covered someDocker containers worth installing early on a new server,and a lot of those projects can live under the same Dockge setup later.

Miniflux gave me the best example of why I wanted this in the first place. Its Compose file referenced a separate .env file, and Dockge picked up both without any extra work. The stack page showed PostgreSQL and Miniflux, while Edit mode put the YAML and environment variables within reach from the browser.

My Compose configuration used MINIFLUX_PORT for the host side of Miniflux’s port mapping, so changing the port meant only updating the corresponding value in .env. My instance originally used 8082.

MINIFLUX_PORT=8085

I changed it in Dockge and clicked Deploy. Under the web interface, Dockge still uses normal Docker Compose behavior. Deploy runs the stack again, recreating services whose configuration changed while leaving the rest alone when it can.

That is exactly what happened here. Miniflux was recreated with the new port mapping while PostgreSQL stayed up and healthy. Dockge’s live output showed the operation progressing in the browser, port 8082 stopped answering, and port 8085 brought up the same Miniflux installation with all of my existing data still there.

That’s the kind of edit that used to send me into SSH. But Dockge condensed the whole little pilgrimage into the page where I was already looking at the stack.

dockge-updating-gitea-compose-stack

The other controls were just as straightforward. Uptime Kuma gave me Start, Stop, and Restart operations that did exactly what I expected. Updating Gitea pulled the configured images and brought both the database and app back up from the same page.

If you keep adding self-hosted apps to spare hardware, those little maintenance jobs add up fast, especially after turning an older PC into an always-on server.

Dockge hasn’t evicted SSH from my server

The files under the hood are still ordinary Compose files

docker-compose-verifying-miniflux-port-change.

After changing Miniflux to port 8085, I went back through SSH because I wanted to see what Dockge had actually done.

The .env file under /opt/stacks/miniflux now contained MINIFLUX_PORT=8085 , and running docker compose ps from the same directory showed the containers running with the new 8085 to 8080 port mapping.

That ended up being one of my favorite parts of the test. Dockge writes your changes back to the actual stack files on disk and keeps them usable with regular Docker Compose commands. The project’s documentation even jokes that it won’t “kidnap” your Compose files, which is a pretty good way to describe the design.

So, I can use the browser when it saves me time, then fall straight back to the CLI whenever I want to inspect, troubleshoot, or change something by hand. I’m still looking at the same Compose files either way, with no proprietary project format wedged into the middle and no weird translation layer to untangle later.

SSH still has plenty of work to do on this server. If Docker itself won’t start, permissions get mangled, networking goes off the rails, storage fills up, or Debian needs maintenance, Dockge can’t fix the layer it depends on. If you bounce between a lot of servers, a dedicated SSH toolkit like OmnySSH that also covers monitoring and file transfers might serve you better.

Dockge also sticks pretty closely to Compose instead of trying to become a cockpit for every Docker resource on the machine. The project points users toward Portainer or the Docker CLI when they need broader container management. I actually like that restraint, but it does mean you should know whether you want a Compose dashboard or a full-blown Docker control panel before you install it.

A smaller ritual is still an upgrade

I’m not uninstalling OpenSSH or implying that a browser can run an entire Linux server for me. Dockge just gives repetitive Compose work a much better front door.

My YAML still lives on the server, Docker Compose still reads it, and the terminal is one PowerShell command away whenever I need it. I just don’t have to drop into SSH because one environment variable changed. After one evening with this workflow, my old routine already looks needlessly baroque.

Leave a Reply

Your email address will not be published. Required fields are marked *