The current implementation of service detection focusses on containers.
However when using Docker Swarm this limits automatic service detection
to the containers running on the same node as homepage (or
docker-proxy). Detecting other services in the docker swarm is with the
current implementation not possible.
This commit fixes this by getting the appropriate labels from services
when Docker swarm is configured in the config file. This ensures that
the appropriate labels are gathered from the service definition instead
of the container definiation, thus allowing for automatic service
detection for the entire Docker swarm. Please note that in order for
this to function the homepage (or dockerproxy) should be running on a
manager node. Only the manager node is able to gather all the relevant
service labels.
Fixes: #752, #970, #955, #1255, #1045, #1496
* Add support for pfSense API
* Fix linting issues
* remove a line
* rename cpu to load in default block
* Re-order container blocks to ensure defaults show
* clean up
* Add support for multiple volumes as well as defining the volume you want to track
* QNAP widget syntax corrections, translate invalid
---------
Co-authored-by: shamoon <4887959+shamoon@users.noreply.github.com>
* Add support for pfSense API
* Fix linting issues
* remove a line
* rename cpu to load in default block
* Re-order container blocks to ensure defaults show
* simplify pfsense widget, rename, limit blocks
---------
Co-authored-by: shamoon <4887959+shamoon@users.noreply.github.com>
* Only environment variables starting with HOMEPAGE_VAR_ and HOMEPAGE_FILE_
are supported
* The value of env var HOMEPAGE_VAR_XXX will replace {{HOMEPAGE_VAR_XXX}}
in any config
* The value of env var HOMEPAGE_FILE_XXX must be a file path, the contents
of which will be used to replace {{HOMEPAGE_FILE_XXX}} in any config
* If a substituted value contains a variable reference it may also be
replaced, but the behavior is non-deterministic
External services define a kubernetes service pointing to an application
hosted outside of the cluster. These services should not attempt to get
the status based on kubernetes deployments. The new boolean flag disables
this functionality. Since this is an edge case, the default value is "false".
Dan5py
shamoon
nsankbeil
Matteo Bossi
Cyb3rzombie
RoyK
Jon
Georges-Antoine Assi
Shaun Walker
David O'Trakoun
Andre
Georges-Antoine Assi
Sebastian Osiński
Kassim Benhaddad
Nina Morgan
Mirek Szajowski
Mathias R
Jason Fischer
James Wynn