Implement a Streaming VMStat Alert

Write a command-line program that reads a continuous stream of `vmstat`-style output from standard input and raises an alert based on a sliding time window. The program is invoked as: `myscript <metric_name> <threshold> <max_exceed_count> <window_seconds>` Example: `vmstat 1 | myscript us 100 50 300` Requirements: 1. The input begins with a header row containing metric names, such as `us`, `sy`, `id`, and so on. 2. Each subsequent row is one sample collected once per second. 3. Monitor the metric named by `metric_name`. 4. A sample is considered abnormal if the metric value is greater than `threshold`. 5. Maintain a sliding window covering only the most recent `window_seconds` samples. 6. If the number of abnormal samples in the current window becomes greater than `max_exceed_count`, print the abnormal samples from that window. 7. The program should process the stream online without storing unbounded history. Assume the input is whitespace-delimited and that the target metric always exists in the header.