Do not schedule node temination if it was turned offline for unknown reason

The plugin rightfully discards nodes offline for node monitor failures. It does not discard nodes that are offline by user request. Every other cause is a middle ground that caused the nodes to be discarded. There are situation when offline cause is used to keep the node idle and therefore not discarded so it should treat all user causes that are not created by monitors so that it will not be discarded.

`JCloudsRetentionStrategy` put the nodes in pending deletion state. The offline cause filtering is different in `JCloudsCleanupThread` (that essentially just speeds things up not waiting for the machines to be put in pending deletion state). Either way, this needs to be unified, perhaps extended for all `MonitorOfflineCause`s and `OfflineCause.ChannelTermination` perhaps.