luminol

Supports multiple anomaly detection (e.g., bitmap_detector, derivative_detector) and correlation algorithms (e.g., cross_correlator) with customizable parameters, allowing tailored analysis for specific use cases.

Does not rely on predefined value thresholds; instead, it assigns anomaly scores to each data point, enabling flexible identification of unusual patterns without manual cutoff settings.

Facilitates automated investigative logic by correlating anomalies with other metrics (e.g., GC, IO, CPU) to rank potential root causes, as highlighted in the library's philosophy for systems monitoring.

Accepts time series data as CSV file paths, dictionaries, or TimeSeries objects, making integration easy with various data sources without extensive preprocessing.