The duplicate-delivery result in a second, non-money domain: bookings (planned, not run)
Not run. Would count two bookings of one slot from the booking server's log, two agents per request, a bare arm against an arm guarded on the slot; the booking server has no fault injection yet.
PlannedDoes it work?
- Designed for
- To test whether the two-agent duplicate-delivery result, measured on refunds, holds in a non-money domain: two bookings of one slot, counted from the booking server's log.
What was tested
We planned a replication of the two-agent duplicate-delivery series on the project’s booking starter service: one request, two agents, a bare arm against an arm guarded on the booking slot, with harm counted as two bookings of one slot in the server’s log. An issue-tracker duplicate-close case was left out, because its harm spans objects. The plan was deferred twice and has not run.
What we observed
No data. The duplicate-delivery result rests on one test domain: refunds, on a simulator in Stripe’s API shape and a five-pair Stripe sandbox pilot.
What this means for Heimdall
The two-agent duplicate result has been measured only on refunds. It bears on the executor’s repeat fence, which the layer states for any domain.
Limits
No harness, scenario, preregistration or data. Until runs exist, the result is not claimed outside refunds.
What remains open
Add fault injection to the booking server, preregister sample and harm rule, and run bare against guarded, or record a decision to drop it.