Outliers and Change Detection
Outlier detection algorithms spot abnormal behaviour in your time series (e.g. sales on Mondays used to be $100k and this Monday it's $150k ) and make sure that only meaningful outliers would be reported, so that you would not be bothered with false alerts.
Change detection algorithms analyze the patterns of your time series, searches for changes in the underlying process and alerts when a important change was found (e.g. sales used to be stable and now started decreasing, the moment when sales started decreasing will be alerted as a change point).
Outlier and change detection are useful for monitoring use cases such as: ETL (processing time, table partition size, rows count, etc.), infrastructure monitoring (CPU and RAM load, processed requests, etc.), sales (numbers of product purchases over time), fraud detection etc.