version: '3'
services:
whoami:
image: jwilder/whoami
ports:
- 8080:8000
disrupt:
image: blamebutton/disrupt
volumes:
# Mount the Docker socket, this is required for interacting with the Docker API.
- /var/run/docker.sock:/var/run/docker.sock
deploy:
placement:
constraints:
# Disrupt needs permission to update services, only managers are allowed to do that.
- node.role == managerFor more examples, check out the examples folder
Note: Docker Swarm does not support running locally built images, so neither does Disrupt.
Configuration of Disrupt is done through environment variables. This is done to make the deployment of Disrupt environment-agnostic.
Below is a list of environment variables available to be configured.
| Name | Type | Options |
|---|---|---|
| UPDATE_DELAY | Integer | default = 300 |
| NOTIFICATION_URL | String | Any Apprise compatible URL |
Disrupt uses the Apprise notification library for Python. Check out their documentation for more advanced usages.
Here is a small list of notification providers that Apprise supports:
- Slack
- Discord
- Telegram
- PushBullet
- Dbus
- Custom notifications to a given URL in JSON or XML format
Feel free to make a feature request or if you have Python experience; pull requests are welcome too!
Did you mount the Docker socket to the container? Check out the example if you want to know how.
In the special case that you are accessing Docker over TCP, you should place Disrupt in the
same network as your TCP socket. Using a Docker socket proxy (like
docker-socket-proxy, or
sockguard) is recommended for enhanced security
in this case. You could then configure the proxy to only allow GET requests for service info,
for example. That way, if the Disrupt container gets compromised it can't do any harm to the
cluster in the form of modifications/destructive instructions.
If you're having trouble getting Disrupt to work, we have a Discord for support and questions.