Description of problem:

The ServiceAccounts for both in-cluster and UWM alertmanager set autoMountServiceAccountToken: true. 
This should be improved and set at the pod level. Hence this will require a change in prometheus-operator and its configuration of Alertmanager pods.

A similar change for Prometheus pods was implemented in https://github.com/prometheus-operator/prometheus-operator/pull/4514. 

https://github.com/openshift/cluster-monitoring-operator/blob/7702f6c7d6e1409dea9197e63dafcb0decbe60b9/assets/alertmanager-user-workload/service-account.yaml#L2

https://github.com/openshift/cluster-monitoring-operator/blob/7702f6c7d6e1409dea9197e63dafcb0decbe60b9/assets/alertmanager/service-account.yaml#L2