ML.NET for Memory Leak detection?

The README here (https://github.com/dotnet/machinelearning-samples/blob/master/samples/fsharp/getting-started/SpikeDetection_ShampooSales/README.md) mentions detecting memory leaks are one of the scenarios that anomaly detection can be used for. I tried using the IID Change point Detection algorithm per the tutorial but it told me the point prior to the data changing. I need it to analyze the trend and detect if is normal or trending positive. The memory data trending positive would be indicative of a memory leak. Can you suggest a way I can use ML.NET to analyze a time series and let me know if it shows a slowly growing positive trend?
Thanks in advance.