Many operators have a dilemma on their hands. To find the lead indicators of network failures requires analysis of events leading up to the failure. But many operators are so snowed under that they are too busy rectifying the network to have time to analyse the causes.
Not only that, but many actually rely on lag indicators (ie customer calls, which often don’t get raised until a significant period of time after the event occurred). Naturally, customer satisfaction is already diminishing by the time a lag indicator appears.
Somewhere in between is the traditional alarm list provided by traditional alarm management suites. By the time a near-real-tine alarm / event filters through, the customer satisfaction and / or SLA countdown clocks have already started. They’e still lag indicators.
This is why advanced predictive tools are becoming so highly sought after and why machine learning is being applied to the problem of failure analysis and predictive events.