Currently, procesors are sending data to recovery when message's processing returns 'failed' status.

Make an ENV 

PROCESSOR_CONTROLLER_RECOVERY_STATUS

to determine when to send data to recovery.

So we will be able to send nothing to all depending on situation with kafka

comma-separated values can be:

• all
• new
• failed
• etc. (list of statuses will be available in 'tower_analytics_report/models_tenants/message.py'