Support "Soft" stopping of processors and queue draining before undeploy

[davetorok]

To prevent in-flight data loss when undeploying / redeploying templates:

• provide a new section in the "undeploy" which takes a list of processgroup-processors (e.g. "Get Kafka 123", "Get RabbitMQ 456", "ExecuteSql 789" to stop
• iterate through connections and wait until all queues are drained to zero, in addition to the current waiting until all processor threads are stopped
• then proceed with the undeploy / delete process group etc.