Multi-target pattern: /probe endpoint cannot handle array of targets from ServiceMonitor

What did you do?
I created a service monitor to use multi target pattern by feeding an array of targets to the targets params.

apiVersion: monitoring.coreos.com/v1
kind: ServiceMonitor
metadata:
  annotations:
    meta.helm.sh/release-name: prometheus-postgres-exporter
    meta.helm.sh/release-namespace: monitoring
  labels:
    app.kubernetes.io/managed-by: Helm
  name: prometheus-postgres-exporter
  namespace: monitoring
spec:
  endpoints:
  - targetPort: http
    path: /probe
    params: 
      targets:
        - postgres:password@svc-postgres.ns-app1.svc.cluster.local:5432/app_db?sslmode=disable
        - postgres:password@svc-postgres.ns-app2.svc.cluster.local:5432/app_db?sslmode=disable
        - postgres:password@svc-postgres.ns-app3.svc.cluster.local:5432/app_db?sslmode=disable
        - postgres:password@svc-postgres.ns-app4.svc.cluster.local:5432/app_db?sslmode=disable
    relabelings:
      # - sourceLabels: [__address__]
      #   targetLabel: __param_target
      - sourceLabels: [__param_target]
        targetLabel: instance
      - targetLabel: __address__
        replacement: prometheus-postgres-exporter:80
    metricRelabelings:
    - sourceLabels: [instance]
      regex: (.*\.(\w+-\w+).*)
      targetLabel: shortname
      replacement: ${2}
      action: replace
  jobLabel: prometheus-postgres-exporter
  namespaceSelector:
    matchNames:
    - monitoring
  selector:
    matchLabels:
      app: prometheus-postgres-exporter
      release: prometheus-postgres-exporter

What did you expect to see?
I expected this to work, similarly to blackbox exporter which accepts an array of targets.

What did you see instead? Under which circumstances?
/probe sends a 400 error bad request to prometheus as /probe cannot handle an array of values being sent to it.

Environment

• System information:

  Kubernetes 1.23
  Helm3
  

• postgres_exporter version:

  postgres-exporter:master

• postgres_exporter flags:

/bin/postgres_exporter --extend.query-path=/etc/config.yaml --web.listen-address

• PostgreSQL version:

  PostgreSQL 14.7 (Debian 14.7-1.pgdg110+1)

• Logs:

no relevant logs

As an alternative I am currently running multiple endpoints, each with its own target, but this is hard to maintain (servicemonitor will have many config duplicates) when you have many postgres instances to monitor from a single postgres-exporter deployment.

Start at the /probe endpoint and reproduce the 400 response with the repeated targets shown in the ServiceMonitor configuration. Compare the desired behavior with the blackbox exporter’s array handling, then define and verify how multiple PostgreSQL targets should be processed while preserving the current single-target behavior.