Managing Data Operating System
Also available as:
PDF
loading table of contents...

Timeline Service 2.0 Overview

YARN Timeline Service 2.0 is the next major version of timeline server after versions 1.0 and 1.5. Timeline Service 2.0 uses a distributed writer architecture and a back-end storage that are more scalable than the earlier versions.

YARN Timeline Service 2.0 addresses the following two major challenges associated with the earlier versions of Timeline Server:

Scalability
Timeline Server 1.0 is limited to a single instance of reader or writer and storage. Therefore, it does not scale well beyond small clusters. YARN Timeline Service 2.0 separates data writes from data reads. It uses distributed collectors, essentially one collector for each YARN application. The readers are separate instances that are dedicated to serving queries through REST APIs.

YARN Timeline Service 2.0 uses Apache HBase as the primary backing storage, because Apache HBase scales well to a large size while maintaining good response times for reads and writes.

Usability
YARN Timeline Service 2.0 provides improved usability because it supports aggregation of information at the level of flows or logical group of YARN applications. It supports configuration information and metrics.