In the current implementation of the service config-sync, a secondary node must be online.
So config-sync uses API calls to retrieve the local configuration of the desired sections and send it to the secondary node via API calls.
But if the secondary node in the state is offline or temporarily unreachable via API, the real config-sync can not be done.

Add the ability to resync configured sections when the secondary node is available again if something was changed on the primary node (in configured sections) during this time.
The primary node should re-send the configuration when the secondary node is available, or the secondary node should ask for config from the primary instance when it has a connectivity/online state.

For implementation, it could be some queue where we place the config-sync commands/sections (depending on the mode):

1. If the queue is not empty, we try to check secondary node availability via the API
2. If the secondary node is unreachable, waiting for some timeout and again try point 1
3. If the secondary node is available, send config to the secondary node from the queue of changed sections

Or other options like based on etcd solutions.