Recently, due to a problem with the computer, there is no update. Today, I will mainly introduce the use of Flink's web ui. Let's submit a job first (my cluster is on yarn). The command is as follows:
flink run -m yarn-cluster -c flink.window.FlinkWindowDemo -yn 2 -ys 4 -nm hbase -yjm 2048 -ytm 2048 -p 9 /home/jason/bigdata/jar/flink-1.0-SNAPSHOT.jar -d
Wait a while after submission, because Flink's UI response is still a bit slow.The home page is as follows, you can see that the entire UI interface is still relatively simple, of course Ali will be optimized later.
First look at the menu bar on the left, we will introduce one by one
1.Overview, you can see that the number of tm applications for our job, the number of task slots, and the number of available task slots are displayed in the middle, and the total number of jobs, running jobs, and completed and cancelled jobs are displayed on the right. And the number of failed jobs, the following is a list of running jobs, and a list of completed jobs
2, Then we click on the running job to enter, and we can see the detailed information of the job we just submitted, as shown in the figure below.
You can see that the above is the DAG diagram of the job, and the following are subtasks, task monitoring, watermarks, accumulators, checkpoints, and being suppressed. The specifics inside are not explained in detail. You can click in and see. We set the task concurrency, the number of data received, and other information.
3. The completed job, this needs to be configured with history-server and started, and then you can see it. It will not be explained in detail here, but will be introduced later.
4, tm let's click directly in and take a look,