What does the FailedInvocation metric refer to for a custom EventBus rule?

We have a custom EventBus "A" with a rule to forward to a global EventBus "B" in another AWS account. A lambda makes a request to EventBus "A" with data that should eventually flow to EventBus "B".

Currently, EventBus "B" does not have EventBus "A" allow listed as we are using a testing environment for EventBus "A" infrastructure. This is fine for our testing purposes. However we're seeing that "FailedInvocation" metric in the AWS console under the "Monitoring tab" of the rule is being incremented as the lambda is triggered and we're confused on what this refers to.

Does this "FailedInvocation" metric refer to "lambda-to-EventBus-A" data flow or "EventBus-A-to-EventBus-B" data flow?

The former if true is unexpected for us and the latter if true would be expected.