Why Docker is one row in Activity Monitor
Every engine on a Mac runs containers inside one Linux virtual machine, and macOS sees the machine, not what is inside it. Activity Monitor shows the engine at eight gigabytes and nothing more. Which container grew, which is busy and which holds the port are things only the engine knows. A Mac with a few projects’ worth of containers runs short of memory, and the cause is one row you cannot open.
The machine also keeps the memory it was given whether containers are running or not. Vitals shows what is inside it. Capping the machine is done in Docker’s own settings.
What Vitals shows for each container
Each container is listed on its own with its memory against its limit, its CPU, what it is sending and receiving, the ports it publishes and its image. Everything is measured inside the engine’s virtual machine, so it is part of the engine’s memory and not counted on top of it.
Cards above the list add up the containers’ memory and CPU, name the container closest to its memory limit, and count the published ports.
The Containers section appears while an engine is running and goes away when it stops. The dropdown’s Developer tab has it too, and the dropdown’s Overview gets a Containers tile.
Engines it works with
Docker Desktop, OrbStack, Colima, Lighter, Rancher Desktop and Podman are recognised. Vitals also follows whichever engine your docker command is set to use, so an engine it has never heard of works if it registered a Docker context.
The docker command does not need to be installed. Vitals talks to the engine directly.
Stop, and nothing else
Vitals is read-only apart from Stop. Click Stop on a container, in the window or the dropdown, and Vitals asks first. It then stops the container the way docker stop does, giving it ten seconds to shut down before it is killed. It never starts, restarts or removes anything.
Compose containers under their project
Containers started by docker compose carry their project and service name and the folder they ran from. In the Projects section they sit under that project beside its dev server and tests, so a project’s whole footprint is one row. Quit the project and its containers stop with it.
How to use it
- 1
Start your engine
The Containers section appears in the Developer tab while an engine is running.
- 2
Read a container
Memory against its limit, CPU, network, published ports, image, and its Compose service and folder.
- 3
Stop one
Click Stop, in the window or the dropdown. Vitals asks first, then gives the container ten seconds before it is killed.
- 4
Follow the project
In Projects, a project’s Compose containers sit under it with its dev server.
The details
- Engines
- Docker Desktop, OrbStack, Colima, Lighter, Rancher Desktop and Podman
- Per container
- Memory against its limit, CPU, network, published ports, image
- Compose
- Project, service name and folder; listed under the project
- Stop
- As docker stop does: ten seconds, then killed, after a confirmation
- Never
- Starts, restarts or removes a container
- In Terminal
- vitals containers
Questions
Which Docker engines does Vitals support?
Docker Desktop, OrbStack, Colima, Lighter, Rancher Desktop and Podman, and whichever engine your docker command is set to use.
Why does Activity Monitor show Docker as one process?
On a Mac every engine runs containers inside one Linux virtual machine, and macOS sees only the machine. Vitals asks the engine what is inside.
Can Vitals start or remove containers?
No. It is read-only except for Stop, which it does only when you ask and after a confirmation.
Does it need the docker command installed?
No. Vitals talks to the engine directly.
Why does Docker use so much memory with nothing running?
The virtual machine keeps the memory it was given whether containers run or not. Capping it is in Docker’s own settings.
The same feature on the app's own site: vitalsmac.com/features/docker