You are working with a dataflow that is processing large volumes of data from multiple sources. The dataflow is experiencing performance issues due to data skew. Describe how you would identify data skew in the dataflow and what steps you would take to resolve it. Consider both data-level and configuration-level optimizations.