This article announces over 30 new features for Amazon EMR on EC2 clusters, focusing on improved scalability, resiliency, and availability for big data processing workloads.

• Enhanced Spot Instance interruption handling prevents cluster scaling operations from getting stuck
• Recommissioned nodes now accept Spark tasks within seconds instead of 60 minutes
• Improved YARN exclude file management reduces scale-down failures from disk over-utilization
• Local storage file system more resilient to instance reconfigurations and EBS volume remounting
• Fixed timing sequence issues reducing Kerberos cluster startup time by up to 200%
• Log management daemon automatically restarts if interrupted, preventing disk over-utilization
• Log rotation enabled for long-running clusters to prevent disk space exhaustion
• Expanded log monitoring covers additional log folders for better cluster stability

These enhancements make Amazon EMR clusters more stable and efficient for long-running and large-scale data processing workloads.