Description of problem:
Whenever a service monitor references an invalid secret or configmap's key, the prometheus operator wouldn't update the Prometheus configuration. It shouldn't be a big issue for the infra Prometheus because we pretty control what goes in but it's more problematic for user workload monitoring (basically a bad service monitor can DoS the service).

Version-Release number of selected component (if applicable):
4.6

How reproducible:
Always

Steps to Reproduce:
1. Enable user workload monitoring
2. Create a secret + a service monitor that references this secret but with an invalid key
apiVersion: v1
data: {}
kind: Secret
metadata:
  name: demo
  namespace: default
type: Opaque

apiVersion: monitoring.coreos.com/v1
kind: ServiceMonitor
metadata:
  name: demo
  namespace: default
spec:
  endpoints:
  - port: web
    bearerTokenSecret:
      key: missing
      name: demo
  selector:
    matchLabels:
      app: demo

3. Create a valid service monitor

apiVersion: monitoring.coreos.com/v1
kind: ServiceMonitor
metadata:
  name: demo2
  namespace: default
spec:
  endpoints:
  - port: web
  selector:
    matchLabels:
      app: demo2

Actual results:
The second service monitor isn't present in the Prometheus configuration.

Expected results:
The second service monitor should be present in the Prometheus configuration.

Additional info:
https://github.com/prometheus-operator/prometheus-operator/issues/3327