Before this, we didn’t have a `healthcheck` at all, but instead there
was simply a check to see if the container was running.
Checking whether the container is running has been moved to Deploy stage
and now occurs at the end of the deployment process.
And between deployment and “init” there is now a correct health check.
Applications are not required to support healthcheck at all, so it is
only checked if `['State']['Health']['Status']` is present.
Without a timeout, the timeout must be set by the application itself, as
it is usually done for Docker containers healthcheck.
During a healthcheck, an application, for example, can now install some
of its own packages or do something other with its docker container.
It should not communicate with the Nextcloud itself at this
stage(healthcheck), because application is not considered enabled.
---------
Signed-off-by: Alexander Piskun <bigcat88@icloud.com>
Signed-off-by: Andrey Borysenko <andrey18106x@gmail.com>
Co-authored-by: Andrey Borysenko <andrey18106x@gmail.com>
