watchdockContainer monitoring
GitHub
Open source · Self-hosted · MIT

A tiny watchdog for your Docker containers.

watchdock runs as a single ~11 MB container, watches every other container on your machines — local or remote over SSH — and pushes a notification to your phone via ntfy when one crashes, goes unhealthy, or recovers.

Try the live demo Get started

$ git clone https://github.com/cobanov/watchdock && cd watchdock && docker compose up -d --build

The watchdock dashboard: containers across hosts, with per-host tables showing image, status and uptime The watchdock dashboard in dark mode: containers across hosts, with per-host tables showing image, status and uptime

Everything you need, nothing to babysit

Point it at a Docker socket and forget about it. watchdock keeps a live state map of your fleet and turns transitions into notifications.

Push to your phone

Crashes, failing healthchecks and recoveries land on your phone in seconds — through public ntfy.sh or your own ntfy server, with token auth supported.

Your whole fleet, grouped

Group by host, status, health or image, filter as you type, and click any container for its full details. States are colour-coded and update live, and every start, stop and crash is kept in the event log.

Knows a crash from a stop

Exit codes and healthcheck transitions are tracked per container. A manual docker stop is never mistaken for a crash.

Remote hosts over SSH

Add a host from the UI over SSH keys or ssh-agent — nothing to install on the other side. Import and export hosts as JSON, and watchdock picks up your ~/.ssh/config too.

Quiet by design

Notifications fire on state transitions only, rate-limited per container and per type — a crash loop won't flood your phone. An ignore list mutes the rest.

Tiny and config-free

A single ~11 MB static Go binary with the UI embedded. No config files to edit — everything's set in the UI and saved to a Docker volume, with light and dark themes.

Five signals, your rules

Each notification type has its own toggle, straight from the web UI. Want only the bad news? Keep unhealthy and crashed. Want the full play-by-play, starts and stops included? Switch everything on.

  • Unhealthy — a healthcheck starts failing
  • Crashed — died with a non-zero exit code
  • Recovered — back healthy, or back up
  • Stopped — clean exits and manual stops
  • Started — a container starts or restarts
Alert rules
Fires only on state transitions, rate-limited per container.
Unhealthy containers
A container's healthcheck starts failing
Urgent
Crashed containers
Died with a non-zero exit code
High
Recoveries
Back to healthy, or back up after a crash
Default
Stopped containers
Clean exits and manual stops
Default
Started containers
A container starts, including restarts
Default

Running in two minutes

Works anywhere Docker Desktop or the Docker daemon runs — macOS, Windows (WSL2), Linux.

  1. 1

    Run it

    Clone the repo and bring it up with compose. The image builds locally in seconds.

  2. 2

    Pick a topic

    Open http://localhost:9622, set an ntfy topic, hit Save, then Send test notification. Pick something unguessable — anyone who knows the topic can read it.

  3. 3

    Subscribe on your phone

    Install the ntfy app (iOS / Android) and subscribe to the same topic. That's it.

# grab it
git clone https://github.com/cobanov/watchdock.git
cd watchdock

# run it
docker compose up -d --build

# open the dashboard
open http://localhost:9622