Anyone who has deployed machine learning at scale can tell you about those mornings when they wake up to find at least a dozen ML pipelines failing. However, few solutions exist to address this pain. Why Labs has released a Python and Java implementation of WhyLogs, a statistical logging library that helps teams monitor and debug data engineer and machine learning pipelines. WhyLogs enables users to detect data quality issues and/or data drift/feature shift by calculating approximate statistics and data sketches that represent any changes to the statistical properties of model inputs and outputs.