In our project we have healthcheck scripts on every server which monitor cpu/ram usage and queue for processes. All of these have multiple thresholds and alerts are sent via sendmail for support teams when maximum threshold are near.

which is similar to what you have described