Skip to content

Managing Stacks

github-actions[bot] edited this page Nov 9, 2026 · 10 revisions

Managing Stacks

This page covers all the lifecycle actions you can perform on a stack: bringing it up, stopping it, starting it, taking it down, restarting it, pausing it, and killing it.

Working with multiple stacks? Select more than one stack and run Up, Restart, Pull, Down, or Kill on all of them at once — see Bulk Actions.

Up

What it does: Runs docker compose up -d. Docker creates or recreates containers as needed, applies configuration changes, and leaves everything running in the background.

When to use it: When you want to start a stopped stack, or after editing the compose file and want to apply the changes.

Confirmation step

Before the operation starts, a confirmation modal always appears. It shows:

  • A warning that containers with changed configuration will be recreated.
  • A list of every service and its image. Services using unpinned images (:latest or no tag) are marked with a ⚠ warning — unless their pull_policy rules out fetching (never, or missing for an image already on disk). A service's pull_policy, if set, is shown next to it.
  • An Optional profiles section — if your compose file defines profiles, checkboxes appear here so you can select which ones to activate. Only services tagged with a selected profile will start. Services without any profile tag always start regardless.
  • A Jobs list, if the file declares jobs (Docker Compose 5.6.0+). Up never starts jobs; trigger them from ⋮ → Run (see Running Commands). If a job with a schedule is active (it has no profile, or one of its profiles is selected), a red warning explains that Compose will refuse to start the stack, because scheduled jobs are not supported yet. The warning follows the profile checkboxes, and it does not block Up: the Compose CLI has the final say.

Click Up to proceed or Cancel to abort.

Progress modal

A progress modal opens and streams the output of docker compose up in real time. The log is color-coded by event type. When the operation finishes:

  • ✓ Up complete — all containers started successfully
  • ✗ Up failed — an error occurred; the full error is shown below the status

If Docker Compose prints warnings, for example about attributes that have no effect outside Swarm mode (Compose 5.6.0 warns about deploy.update_config, credential_spec, ports[].mode: host and others), they are collected into a Docker Compose reported warnings box above the log, as plain messages. A warning does not fail the operation. The Pull progress modal does the same.

Click Close when done. If the operation is still running, click Cancel to abort it. Alternatively, click Run in Background (or close the modal with its ✕) to close it and keep the operation running — see Background Tasks.


Start

What it does: Runs docker compose start. Starts existing stopped containers without recreating them. Config changes in the compose file are not applied.

When to use it: When you stopped a stack temporarily and want to resume it exactly as it was.

No confirmation modal. The button triggers the command directly and the row updates as containers come online.


Stop

What it does: Runs docker compose stop. Sends SIGTERM to all running containers and waits for them to exit gracefully. Containers are preserved — they can be started again with Start.

When to use it: When you want to pause a stack temporarily without losing container state.

No confirmation modal. The dashboard updates as containers stop.


Down

What it does: Runs docker compose down. Stops and removes all containers for this stack. The stack disappears from the Running section. Networks created by the stack are also removed. Volumes and images are not removed by default.

When to use it: When you want to tear down a stack completely and remove its containers.

Confirmation dialog

A confirmation modal appears:

Running docker compose down will stop and remove all containers for [stack name].
The stack will disappear from this list.

If any of the stack's networks are also used by other running stacks, a warning is shown listing which stacks share those networks. Proceeding will remove the network, which may affect the other stacks.

Click Down (remove) to proceed or Cancel to abort.

After Down completes, the stack may reappear in the Stopped / offline stacks section if its compose file still exists on disk.

When containers are left behind

After Down (and Kill), the plugin checks what Docker still lists for the stack. If anything is left, the dialog stays open and names the leftover containers — also when compose itself failed, for example because the stack's folder was deleted. Click Force remove to delete them by their project label; this works without the compose file. Down and Kill follow the same rule: a stack is gone when no containers are left, whatever the command reported. If the command failed but nothing is left, the dialog closes and a warning toast shows the error. If Force remove fails, the dialog shows why.

If Docker lists containers it cannot open, their records are damaged (usually after a crash or power loss), and Docker can neither start nor remove them. The dialog shows the exact fix for your setup and a Repair button. Repair stops Docker, deletes the damaged records and starts Docker again. While Docker is stopped, every container on that daemon stops, and containers without a restart policy stay stopped afterwards. Repair is Docker-only; Podman stores its state differently.


Restart

What it does: Runs docker compose restart. Stops and then starts all containers without recreating them. Config changes in the compose file are not applied.

When to use it: When a service has become unresponsive and you want to cycle it without changing anything else.

Found in the ⋮ more menu. Disabled for stopped or unknown stacks.


Pause / Unpause

What it does:

  • Pause — Freezes all running containers in the stack. CPU usage drops to zero. Memory is retained.
  • Unpause — Resumes frozen containers from where they left off.

When to use it: When you need to temporarily freeze a stack (e.g., to take a consistent snapshot) without stopping it.

Found in the ⋮ more menu. Pause is shown when the stack is running; Unpause is shown when the stack is paused (status label turns blue).

While a stack is paused, Up is greyed out (unpause it first), but Stop and Down still work: the stack's paused containers are unpaused for you first, since Podman refuses to stop a paused container.


Kill

What it does: Runs docker compose kill. Sends SIGKILL to all containers, terminating them immediately. Unlike Stop, containers have no chance to clean up or flush buffers.

When to use it: Only when Stop is not working — for example, when a container is hung and not responding to graceful shutdown signals.

Found in the ⋮ more menu. This is a danger action.

Confirmation dialog

Running docker compose kill sends SIGKILL to all containers in [stack name], forcefully terminating them immediately.
Unlike Stop, processes have no chance to clean up. Use only when Stop does not respond.

Click Kill all containers (danger button) or Cancel.

If containers remain after Kill, the dialog lists them the same way as Down — see When containers are left behind.


Scale

What it does: Adjusts the number of running replicas for one or more services (docker compose up --scale).

When to use it: When you need multiple instances of a stateless service for load distribution or testing.

Found in the ⋮ more menu. Only available for running stacks. See Scaling Services for the full walkthrough.


Notifications and errors

Start, Stop, Restart, Pause and Unpause — and the per-service Start, Stop and Restart buttons — confirm the result with a short notification in the corner: green when it worked, red with Docker's error message when it didn't.

If an action fails, a red error alert also appears directly below the stack row with the error message from Docker. The alert disappears when you close it or trigger another action.

Clone this wiki locally