v2.2.0 is now available after a few weeks of beta testing. Thanks to everyone how has tested the beta and contributed code!
This release is huge and here are all the changes from v2.1.10:
Agent!
- Support for Agents. This is probably the most requested feature to date. With the Agent you don't need to open the Docker socket on remote hosts to manage them. Instead, just install the Agent, either as a container or as a system service which means no exposed ports and better security. To do this, go to Hosts and click on "Add Host". DockMon now defaults to the Agent as the recommended option and it will show you installation instructions in the UI.
- Migrations from current remote hosts is supported as well so you don't have to remove an existing remote host first, just install the Agent and DockMon will handle the migration
- Easy Agent updates. The Agent support self-updates which makes running the latest version much easier. The container version updates just like any other container whereas the system service version updates via the Host information page.
New features
- Improved charts on the host and container drawer/modals with axis descriptions, dynamic Y axis and hover tooltip (#86) (thanks @marvinvr)
- Mobile Optimizations (#86) (thanks @marvinvr)
- Ability to add a "Update Now" link/button using a new {ACTION_URL} variable so you can update a container directly from a notification. See #16 for more details
- Ignore events from containers using wildcards (#34)
- Option to ignore container updates completely, new configuration in Settings -> Container Updates (#85)
- Alert Clear Delay to allow for temporary issues to not generate an alert (#99)
- Compose file import, makes it possible to import stacks/compose files from other systems like Portainer, Dockge etc. (beta feature)
- Derived tags for agent containers can now be used in alert configurations (#88)
- Added support for Docker Compose .build
Other changes
- Completely rewritten deployment code. This used to be convoluted Python code that was just too complex and meant each capability in Docker Compose has to be replicated in code. Now DockMon is using a the Go Compose SDK for all deployments instead which means deployments will be a lot more reliable.
- A similar refactor was done to the update system which is now also written in Go and the code is shared between DockMon and the Agent.
- Replace -backup- container suffix with -dockmon-backup (#76) (thanks @guilhermenovais)
- The alert system didn't allow you to use compose tags as an option for sending alerts (#88)
- A stopped container will now be stopped immediately after an update is successful (#90)
- Use compose keys for container selection in alert rules. (#99)
- Bypass registry cache for manual update checks (#101)