In the alerts documentation:
For the condition "Cluster is missing an active mongos", it says the alert trigger is "Ops Manager cannot reach a mongos for the cluster." At least one customer was misled into believe that to mean they would be alerted if a mongos router wasn't running. In fact, that alert will only fire if *all* mongoesses are down.
Please if possible change the alert trigger description to
Ops Manager cannot reach any of the mongos for the cluster.