In the current design only active events can be selected in the tester. The tester should permit testing events before they are made active.
Requiring events to be made 'active' during development adds risk to the Tealeaf system as new and raw events must be propagated to the canisters for evaluation. If a processing loop, inefficient advanced code, RegEx, etc., overwhelms the event processors, event evaluation can come to a halt until the event is made inactive again.
Additionally, the current event tester design writes data to the Data connector (and perhaps event bus) when these features are enabled. This writing should not occur, as testing is not actual session activity and should not trigger alerts or effect metrics in the downstream data consumers. At minimum, if an inactive event is permitted to the tested, it should not send anything with a CxConnect context.
How will this idea be used?
Allowing inactive events to be tested in the event tester will protect the Tealeaf system. We have had a few incidents where event development has harmed production event evaluation. |
|
What is your industry? | Healthcare |
What is the idea priority? | Medium |
Note, also, event development in a non-production system has it's own drawbacks, as events and hit attributes may have different internal IDs. In that situation migrating an advanced event to production requires a re-write, to align these IDs. Importing is dangerous as it may introduce duplicate items or overlay items that should not be overlaid.