Content

Target customers: all users. Features released: EMR Doctor provides daily cluster reports. 1. The YARN engine is added. EMR Doctor can analyze the YARN engine to display the resource allocation and usage of clusters in multiple dimensions. This helps improve resource utilization. a. The peak and off-peak hours in which YARN memory resources are used each day can be analyzed. b. The statistics on the number of YARN tasks that are running in different periods of time can be collected to determine the busy and idle periods. c. Basic information about the YARN engine can be displayed. d. Resource usage in queues and resource usage by user can be analyzed. 2. Computing-related items are added for analysis. a. Whether long tails occur in tasks can be detected. b. Fine-grained diagnosis of data skew by stage is supported. c. I/O- and shuffle-related information can be displayed in the job details. 3. The data that is relevant to memory usage and analysis time is optimized for cluster storage analysis.