IllegalStateException: Vert.x instance is closed – alert design to suppress noise from closed-instance misuse

Goal

Define an alerting strategy for Vert.x 4.x that captures genuine lifecycle failures — such as the IllegalStateException: Vert.x instance is closed thrown when any Vertx API is invoked after vertx.close() — while avoiding notification storms caused by repeated misuse of a closed instance.

Constraints and uncertainty

The exception is a clear signal of programming error, yet applications that retry deployment or event-bus calls on a closed instance can generate a flood of identical exceptions. Vert.x provides a HealthChecker SPI and Circuit Breaker for failure thresholds, but the optimal consecutive-failure count for production workloads is undocumented. Separately, the community has not settled whether blocked-event-loop detection should be exposed as a gauge (current blocked count) or a histogram of block durations; the current Micrometer bundle ships a gauge, which limits alerting granularity.

Questions

1. What failure-threshold configuration (e.g., number of consecutive health-check failures) effectively distinguishes a transient spike from a true closed-instance condition without missing real outages?
2. Until the blocked-event-loop metric design is resolved, should alert rules rely on the existing gauge, or is a custom histogram wrapper advisable?
3. Are there recommended patterns for catching and suppressing the closed-instance exception at the framework level rather than at each call site?