Graymail engine restarting due to Graymail Rules update.
There is a 'manual' engine rollback feature - does not work - this is a rules update issue, not an engine update issue.
We would like 2 engines available.
- Current engine version
- + Previous engine version
The same for Rule Updates
- Current rule update
- + Previous rule update
Should the system detect an engine restart 3 times, then the Engine + Rule Updates since the first detection are marked bad.
If the engine has been update in the past 24hrs, use Previous Engine.
Else, use Previous Rule Update to last known good (not marked bad update).
As new Engine Updates or Rule Updates are received, send a copy of a message in pipeline through the Engine+Rule and discard. If successful 3 times, update Engine (if required) + Rule Update and mark both Engine + Rule Update good.
( 3 times - assuming historically any message killed the engine )
System reverted to a good state - automatically, without manual intervention.
