Container Selection¶

Updock provides several mechanisms to control which containers are monitored and updated.

Default Behavior¶

By default, Updock monitors all running containers on the Docker host. This can be changed using labels, flags, or container name arguments.

Enable / Disable Labels¶

Opt-in Mode (`--label-enable`)¶

When --label-enable is set, Updock ignores all containers unless they carry the enable label:

services:
  myapp:
    image: myapp:latest
    labels:
      - "com.updock.enable=true"

Warning

With --label-enable, containers without the label are completely invisible to Updock — they will not be monitored or updated.

Opt-out Mode (default)¶

In the default mode (--monitor-all=true), all containers are monitored. You can exclude specific containers with:

services:
  database:
    image: postgres:16
    labels:
      - "com.updock.enable=false"

Monitor-Only Label¶

To track image updates without applying them, use the monitor-only label:

services:
  production-db:
    image: postgres:16
    labels:
      - "com.updock.monitor-only=true"

Info

Monitor-only containers appear in the Web UI and API with available updates listed, but Updock will never restart or recreate them.

Container Name Arguments¶

You can pass container names directly as positional arguments to limit a run to specific containers:

docker exec updock updock myapp nginx redis

This is particularly useful with --run-once for targeted, on-demand updates.

Disable Containers Flag¶

The --disable-containers flag accepts a comma-separated list of container names to exclude:

docker run -d \
  -e UPDOCK_DISABLE_CONTAINERS="postgres,redis,mongo" \
  updock/updock

These containers will be skipped even if they have com.updock.enable=true.

Scope Filtering¶

The --scope flag limits Updock to containers matching a specific scope label. See the Multiple Instances guide for details.

Stopped & Restarting Containers¶

By default, only running containers are considered.

Flag	Effect
`--include-stopped`	Also monitor containers in `exited` state
`--include-restarting`	Also monitor containers in `restarting` state
`--revive-stopped`	Restart stopped containers after updating their image

Tip

--revive-stopped implies --include-stopped. Use it when you want Updock to bring back containers that were stopped but have a newer image available.

Policy-Based Selection (updock.yml)¶

In addition to labels and flags, containers can be configured in the updock.yml policy file. This is the recommended approach for complex setups:

containers:
  nginx:
    policy: conservative    # assign a named policy
    schedule: "02:00-05:00" # maintenance window

  postgres:
    policy: locked          # never auto-update

  legacy-app:
    ignore: true            # completely invisible to Updock

Containers marked with ignore: true in the policy file are skipped entirely, regardless of labels or flags. See Policies for full details.

Precedence Summary¶

The selection logic is evaluated in order:

Policy file ignore — ignore: true in updock.yml always excludes
Name arguments — if names are provided, only those containers are included
Disable list — container must not be in --disable-containers
Opt-out label — com.updock.disable=true always excludes
Scope filter — container must match --scope (if set)
Label mode — if --label-enable, container must have com.updock.enable=true
State filter — container state must be eligible (running, stopped, restarting)
Policy strategy — strategy: pin prevents auto-update (requires approval)
Maintenance window — schedule in policy file restricts update times