This topic describes how to view job instance details in MaxCompute Studio.
View job instance details
- Use a Logview URL or a local Logview file to view the details about a job instance
in read-only mode.
This is a common method to view details about job instances. You can also use Logview to view the status of tasks that are submitted by other users in other projects. You can view details about a job instance by entering a valid Logview URL in MaxCompute Studio.
In the main menu bar of IntelliJ IDEA, choose Open job detail by logview dialog box, you can copy a valid Logview URL or export a local offline Logview file.. In the
- In Job Explorer, double-click a MaxCompute job instance to view its details.
Job details view
- Execution view: displays the overall information for a job in the form of a directed acyclic graph (DAG). You can view the dependencies between subtasks and detailed execution plans for each subtask.
- Analysis: displays the details about a job by using sequence diagrams, scatter charts, skewed data charts, and long tail distributions.
- History: displays the amount of time consumed in each phase of a job.
- Script: displays the SQL statement and parameter settings that are configured for a job that is submitted.
- Details: displays the details about a job as a table. These details include the subtask list, worker list for each subtask, amount of data that is processed by the workers, execution time, and job status.
- Summary: displays the details about a job in the JSON format.
- Result: displays the execution results of a job.
- Plan: displays the execution plan of a job.
- or : used to expand or collapse views on the left and right panes. These icons allow you to focus on a specific view.
- : used to stop a job that is running. Only the project owner or administrator has the permission to stop running jobs.
- : used to refresh details. The basic information such as status and quota of jobs is automatically refreshed. However, the details views on the right-side pane are not automatically refreshed. If you want to view the up-to-date details, you must manually refresh the views.
- : used to copy Logview URLs.
- : used to generate a Logview URL and access the URL in a web browser.
- : used to save job details as a local file.
- : used to determine whether to enable the auto refresh feature. If this feature is enabled, MaxCompute Studio automatically refreshes all details about a running job on a scheduled basis.
Basic information page
This page displays the basic information about a job, including its ID, creater, status, start time and end time, computing resource usage, input tables, and output tables. The basic information about a running job is automatically refreshed on a scheduled basis.
You can double-click a table name to view the basic information about the table.
The following table describes the meanings of the marked elements in the Execution view.
|①||Click to go to another layer.|
|②||Zoom in and zoom out.|
|③||The dependency table.|
|③||The Fuxi task worker.|
|⑥||Click to display the job dependencies.|
|⑦||The dependencies in the Fuxi Task layer are displayed by default.|
- Fuxi Job layer
Click MaxCompute job to display the Fuxi Job layer. Workers in the Fuxi Job layer include the name, start time, and end time of each Fuxi task.
- Fuxi Task layer
Double-click a Fuxi job worker to go to the Fuxi Task layer.
If multiple Fuxi jobs exist, the Fuxi Task layer of the last Fuxi job is displayed by default. This layer displays the dependencies, input tables, output tables, and partitions of Fuxi tasks. When a job is complete, click the Progress Chart drop-down list to select Input Heat Chart (Record), Output Heat Chart (Record), Input Heat Chart (Byte), Output Heat Chart (Byte), TaskTime Heat Chart, or InstanceTime Heat Chart. The progress chart displays the task progress for the worker. The heat chart uses different colors to distinguish between the heat sources of different workers.A Fuxi task worker shows the following information:
- Instance Count: expressed in the
a/b/cformat. This expression indicates the status of the subtask instances at a specific point in time. The value of a represents the number of running subtask instances, the value of b represents the number of completed subtask instances, and the value of c represents the total number of subtask instances.
- I/O Records: displays the numbers of input and output records at a specific point in time.
- Percentage and orange progress bar: indicates the status of the task. The percentage is obtained based on the analysis of running subtask instances.
- Line that connects subtasks: shows the number of output records. The arrow indicates the direction of a data flow.
- Instance Count: expressed in the
- Operation layer
Double-click a Fuxi task to display the Operation layer.
The Operation layer shows how Fuxi tasks run internally. You can click a worker to view all Operation layer information.
The Details view displays details for SQL data manipulation language (DML) jobs. This view displays the Fuxi task lists and compute worker lists of SQL DML jobs on the compute cluster. In most cases, each job is associated with one or more Fuxi jobs. Each Fuxi job is divided into multiple Fuxi tasks (stages), and each Fuxi task contains multiple Fuxi instances (workers). You can right-click a Fuxi instance to view the standard output, standard errors, and debugging information.
- The Fuxi job tab.
- The Fuxi task list.
- The details and compute worker list of each Fuxi task.
MaxCompute Studio provides the job playback feature. You can view the execution history of a job within 12 seconds in the same manner as playing back a media file. This feature helps you understand the status of a MaxCompute job instance at different time points. You can determine the order and the amount of time to run subtasks and obtain the key path to run the job. This allows you to optimize subtasks that run slowly.
The Timeline view displays the details for the distributed execution of a job in a Gantt chart. You can adjust the display granularity to show all compute workers in a Gantt chart. Gantt charts display the time bottlenecks and long-tail workers of running jobs. You can also use multiple filters to select the key path, the largest data worker, and the worker that consumes the longest amount of time during the job execution.
The Analysis view displays the long-tail workers and skewed data workers of jobs. It also displays worker scatter charts and column charts to help diagnose job execution bottlenecks. Scatter charts and column charts allow you to click a worker to view the details view page and view the details about Fuxi instances.
The Result view displays result information based on the job type and the parameters that are set when you submit the job.
- For SELECT statements in which
odps.sql.select.output.formatis set to HumanReadable, this view displays the result in the text format.
- For SELECT statements that do not specify the output format parameter, this view displays the result in the table format.
- For scripts that are run to export data to a table, this view displays the name of the output table and the link that redirects to the table details.
- For abnormal jobs, this view displays the error details.