MaxCompute job identifiers let you tag jobs at submission time so you can trace them back to their source node, name, and owner in upper-layer scheduling platforms such as DataWorks and DataPhin. MaxCompute parses and records these identifiers from each job's settings.
Parameter identifiers
The following identifiers are supported. Jobs submitted from DataWorks and DataPhin include the platform-set identifiers automatically. Set the user-set identifiers manually when submitting jobs from an SDK or other custom platform.
| Field name | Value | Constraints |
|---|---|---|
EXT_PLATFORM_ID | Job source platform | Use a consistent identifier across all jobs from the same platform. Maximum length: 32 characters. |
EXT_NODE_ID | Scheduling node ID | Identifies the scheduling node in the upper-layer platform. Maximum length: 64 characters. |
EXT_DAGTYPE | Instance run mode | 0: Normal scheduling instance. 1: One-time task. 2: Smoke testing. 3: Data backfill. |
EXT_TASK_ID | Scheduling node instance ID | Identifies the specific running instance of EXT_NODE_ID. |
EXT_NODE_NAME | Scheduling node name | Identifies the name of the scheduling node in the upper-layer platform. Maximum length: 128 characters. |
EXT_NODE_ONDUTY | Scheduling node owner ID | Identifies the owner of the upper-layer scheduling node. Maximum length: 64 characters. |
EXT_NODE_PRIORITY | Scheduling node priority | Identifies the priority of the upper-layer scheduling node. Numeric value stored as a string (settings is a string-to-string map). |
Set parameter identifiers
Set parameter identifiers using SET statements before the SQL query. The following example tags a job with its source platform (SDK), node name (example_task), and owner ID (1234567):
SET EXT_PLATFORM_ID='SDK';
SET EXT_NODE_NAME='example_task';
SET EXT_NODE_ONDUTY='1234567';
sqlquery;Query parameter identifiers
After jobs run, the identifier values are available in two places: the tenant-level Information Schema and the MaxCompute console.
Query the Information Schema
Identifier values are recorded in the tasks_history view of the tenant-level Information Schema. Query this view to associate jobs with their source scheduling nodes for analysis and statistics.
Search in the console
Log on to the MaxCompute console and select a region in the upper-left corner.
In the left navigation pane, choose Observation O&M > Jobs.
The job list displays the ExtPlatformId, ExtNodeId, ExtNodeName, and ExtNodeOnDuty columns for each job. Search by ExtNodeId or ExtNodeName to find a job and identify its source node ID, name, and owner.