It depends. Sometimes proactive - if the system and specific issues are well understood - you have predictive metrics and signs. I do not care if those signs involve measuring the perspiration on the lead engineers forehead. If that person is sweating - something is going wrong but it does not have a name or an incident case number (yet).
Otherwise, you will still be reactive. Anomaly detection is largely reactive. Gonna start a fight - if you act on any metric beyond predictive or trending ones that you know foreshadow a coming issue, then you are probably reactive and not proactive. Everything else is a buzzword or a marketing ploy.
The basic use of SolarWinds is a major step in the right direction. Its about layers!
\m/
