Server Monitoring Suggestions

1) Dispatch an email if the server log is nearing its max. If email messages are configured of course.
2) Handle the error I received gracefully. By gracefully, I mean terminate the service or in the least don't send infinite emails.
3) Make the notification system a little more robust. For example...
Give the option of receiving an email on every failure of a job or to receive an email when it fails and when it is successful the subsequent time.
Allow the failure sensitivity to be configurable. For example, let the user choose after how many failures they should get notified. Or even better have multiple levels which can be dispatched to different email addresses depending on the number of failures. 1 failure - email1, 3 failures - email1,email2, 5 failures - email1,email2,email3 etc...