Description
There are multiple scenarios where it would be useful to look at external metrics to assess service readiness, in addition to the current logic. For instance, by looking at nginx ingress controller metrics, scenarios such as abnormal performance or request count could be identified and the service determined to be unready.
A way to support this would be to add the ability to query Prometheus metrics, for instance somewhat like KEDA's Prometheus scaler does. When designing this implementation, it should be considered if support for Prometheus queries should be directly baked into k8gb or implemented thru an external call, such as:
- externally managed gRPC service, like KEDA's external scalers
- external call to promtool for instance
This implementation should not address single endpoint readiness. That should be done using standard K8s readiness probes even if calling external tools.