Flink REST APIs

Flink has a monitoring API that can be used to query status and statistics of running jobs, as well as recent completed jobs. This monitoring API is used by Apache Flink Dashboard.

The monitoring API is a RESTful API that accepts HTTP GET requests and responds with JSON data. RESTful API is a set of APIs used to log in to the web server. In Flink, web server is a module of JobManager and shares the same process with JobManager. By default, the listening port of web server is 8081. If you want to change the listening port, modify jobmanager.web.port in the flink-conf.yaml file.

The Netty and the Netty Router library are used to handle REST requests and analysis URLs.

RESTful APIs are executed through HTTP requests.

The format of the HTTP requests is http://<JobManager_IP>:<JobManager_Port><Path>

The JobManager_IP indicates the IP address of JobManager, JobManager_Port indicates the listening port of JobManager, and Path indicates the path. For details, see Table 1. For example, http://10.162.181.57:32261/config.

If you want to modify the configuration file flink-conf.yaml of the Flink Client, add to-be-visited IP addresses (separated with commas) in jobmanager.web.allow-access-address and jobmanager.web.access-control-allow-origin parameters.

Table 1 lists all RESTful API paths supported by Flink.

**Table 1** Paths supported by Flink
Path	Description
/config	Some information about the monitoring API and the server setup.
/logout	Some information about the logout.
/overview	Simple summary of the Flink cluster status.
/jobs	IDs of the jobs, grouped by status running, finished, failed, canceled.
/jobmanager/config	the configuration of the jobmanager.
/joboverview	Jobs, grouped by status, each with a small summary of its status.
/joboverview/running	Jobs, grouped by status, each with a small summary of its status. The same as /joboverview, but containing only currently running jobs.
/joboverview/completed	Jobs, grouped by status, each with a small summary of its status. The same as /joboverview, but containing only completed (finished, canceled, or failed) jobs.
/jobs/<jobid>	Summary of one job, listing dataflow plan, status, timestamps of state transitions, aggregate information for each vertex (operator).
/jobs/<jobid>/vertices	Currently the same as /jobs/<jobid>.
/jobs/<jobid>/config	The user-defined execution config used by the job.
/jobs/<jobid>/exceptions	The non-recoverable exceptions that have been observed by the job. The truncated flag defines whether more exceptions occurred, but are not listed, because the response would otherwise get too big.
/jobs/<jobid>/accumulators	The aggregated user accumulators plus job accumulators.
/jobs/<jobid>/checkpoints	checkpoint stats for a job.
/jobs/<jobid>/metrics	a job a list of all available metrics.
/jobs/<jobid>/vertices/<vertexid>	Information about one specific vertex, with a summary for each of its subtasks.
/jobs/<jobid>/vertices/<vertexid>/subtasktimes	This request returns the timestamps for the state transitions of all subtasks of a given vertex. These can be used, for example, to create time-line comparisons between subtasks.
/jobs/<jobid>/vertices/<vertexid>/taskmanagers	TaskManager statistics for one specific vertex. This is an aggregation of subtask statistics returned by /jobs/<jobid>/vertices/<vertexid>.
/jobs/<jobid>/vertices/<vertexid>/accumulators	The aggregated user-defined accumulators, for a specific vertex.
/jobs/<jobid>/vertices/<vertexid>/checkpoints	checkpoint stats for a single job vertex.
/jobs/<jobid>/vertices/<vertexid>/backpressure	back pressure stats for a single job vertex and all its sub tasks.
/jobs/<jobid>/vertices/<vertexid>/metrics	a given task of the values for a set of metrics.
/jobs/<jobid>/vertices/<vertexid>/subtasks/accumulators	Gets all user-defined accumulators for all subtasks of a given vertex. These are the individual accumulators that are returned in aggregated form by the request /jobs/<jobid>/vertices/<vertexid>/accumulators.
/jobs/<jobid>/vertices/<vertexid>/subtasks/<subtasknum>	Summary of the current or latest execution attempt of a specific subtask. See below for a sample.
/jobs/<jobid>/vertices/<vertexid>/subtasks/<subtasknum>/attempts/<attempt>	Summary of a specific execution attempt of a specific subtask. Multiple execution attempts happen in case of failure/recovery.
/jobs/<jobid>/vertices/<vertexid>/subtasks/<subtasknum>/attempts/<attempt>/accumulators	The accumulators collected for one specific subtask during one specific execution attempt (multiple attempts happen in case of failure/recovery).
/jobs/<jobid>/plan	The dataflow plan of a job. The plan is also included in the job summary (/jobs/<jobid>).
/taskmanagers	Some information of the TaskManagers.
/taskmanagers/<taskmanagerid>/metrics	the metrics information of a TaskManager.
/taskmanagers/<taskmanagerid>/log	the log information of a TaskManager.
/taskmanagers/<taskmanagerid>/stdout	the stdout of a TaskManager.
/jobmanager/log	the log of Jobmanager.
/jobmanager/stdout	the stdout of Jobmanager.
/jobmanager/metrics	the metrics of Jobmanager.
/:*	services requests to web frontend's static files, such as HTML, CSS, or JS files.