Problem
Quite often when initially testing in non-production environments, you might want to simulate a service failure by automating a synthetic connection and causing it to fail. This should cause an generated incident based on "drop in number of requests".
Solution
For a service to trigger the drop in number of requests it makes the following assumptions:
The service has been running for at least 5 calls per second for at least 10 minutes.
This shouldn't be an issue with production loads, but can be apparent when testing.
Comments