You determine the cluster size, node specifications, and a range of services, while Yandex Data Proc effortlessly sets up and configures Spark, Hadoop clusters, and additional components. Collaboration is enhanced through the use of Zeppelin notebooks and various web applications via a user interface proxy. You maintain complete control over your cluster with root access for every virtual machine. Moreover, you can install your own software and libraries on active clusters without needing to restart them. Yandex Data Proc employs instance groups to automatically adjust computing resources of compute subclusters in response to CPU usage metrics. Additionally, Data Proc facilitates the creation of managed Hive clusters, which helps minimize the risk of failures and data loss due to metadata issues. This service streamlines the process of constructing ETL pipelines and developing models, as well as managing other iterative operations. Furthermore, the Data Proc operator is natively integrated into Apache Airflow, allowing for seamless orchestration of data workflows. This means that users can leverage the full potential of their data processing capabilities with minimal overhead and maximum efficiency.