I am working on a project where it could be useful to use triggers to update nodes and their associated relationships. In the context of this project there will be around 200.000 nodes and about 150 triggers.
Updates occur at least once per day on the nodes and for each update all triggers start, later maybe also once per hour.
My questions:
-
First, is it smart to use triggers for extensive operations on the database in a live scenario regarding inconsistencies?
-
Is it even traceable/auditable which trigger made a certain change afterwards and is it easy to rollback the change?
-
Is it best practice for live environments in regards of consistency? We are afraid it isn’t and moreover performance will drop.
Alternatively, we think about a classification mechanism which classifies our nodes according to some rules. What is better?