This document discusses the top 5 mistakes to avoid when using Apache Spark in production environments. It identifies the most common issues as: 1) scaling of the architecture to larger datasets and cluster sizes, 2) insufficient memory configurations for drivers and executors, 3) errors in end user code, 4) incompatible dependencies causing classpath issues, and 5) administration and operation-related problems around resource management, security, and configurations. Remedies for each type of issue are provided.