The Timeline Server maintains historical state and provides metrics visibility for YARN applications, similar to the functionality the Job History Server provides for MapReduce.
The Timeline Server provides the following information:
Generic Information about Completed Applications
Generic information includes application-level data such as queue name, user information, information about application attempts, a list of Containers that were run under each application attempt, and information about each Container. Generic data is stored by the ResourceManager in a history store (the default implementation on a file system), and is used by the web UI to display information about completed applications.
Per-Framework Information for Running and Completed Applications
Per-framework information is specific to an application or framework. For example, the Hadoop MapReduce framework can include pieces of information such as the number of map tasks, reduce tasks, counters, etc. Application developers can publish this information to the Timeline Server via the TimelineClient (from within a client), the ApplicationMaster, or the application's Containers. This information can then be queried via REST APIs that enable rendering by application/framework-specific UIs.
The Timeline Server is a stand-alone server daemon that is deployed to a cluster node. It may or may not be co-located with the ResourceManager.
