Predictive maintenance idea to predict servers crashing due to over traffic

As a search, we found a problem in the 'IT' {information technology} sector, in the category of predictive maintenance.

As concerning from some dot net programming engineer we found a problem that whenever a traffic on a particular website increases rapidly and after a certain limit the bandwidth of the server exceeds the server get damaged and comes under the p1 priority which means need to be maintained at the same time as an alternative the server is switched so that traffic might not get lost due to maintenance.

As a solution for this priority problem we, have an idea to create a server interface in between server and the traffic, which continuously monitor the bandwidth limit till 90% of its range and as it researches that level the traffic will be switched to another server as an backup till that time the server comes in its limit to handle the traffic. Also, this interface will generate an alert that server has reached its 90% of the limit and need to be switched so from a predictive maintenance  point of view this can save a lot of time money and also the customer experience at the service the company is providing