AutoETL monitoring and alerting - PolarDB - Alibaba Cloud Documentation Center

AutoETL provides monitoring and alerting to track the status, synchronization latency, and throughput of your search views and ETL links. The lifecycle of an AutoETL link consists of two phases: a full synchronization phase that scans the entire table and an incremental synchronization phase that listens for binlog events. You can use SQL commands to check the ETL status, view metrics in the PolarDB console, and configure alert rules in CloudMonitor.

View ETL status with SQL commands

You can use SQL commands to view the status and error information of search views and ETL links.

View ETL status

View the status of all search views.
```
SHOW SEARCH VIEW STATUS;
```
View all synchronization links created with the dbms_etl stored procedure.
```
CALL dbms_etl.show_sync_link();
```

Result description

In the results, a Status of active indicates the link is running. The Message field provides details such as error messages, synchronization latency, and the number of restarts.

 View Name: view_test
      Type: search
    Status: active
   Message: {"errors": "Unknown database 'db1'", "others": "", "latency": "-1", "jobStatus": "recovering", "restart_times": "1647"}
Created_at: 2026-05-14 12:12:06
Updated_at: 2026-05-14 17:20:42

The following table describes the parameters in the Message field.

Parameter	Description
`errors`	The error messages for the link.
`latency`	The synchronization latency in milliseconds (ms). A value of `-1` indicates that the link has not yet entered the incremental synchronization phase or that there are no incremental writes.
`jobStatus`	The status of the job. For example, `recovering` indicates that the link is restarting.
`restart_times`	The number of times the link has restarted.

View metrics in the console

AutoETL monitors six real-time metrics for job stability, latency, and throughput.

Log on to the PolarDB console. In the left-side navigation pane, click Clusters, select your cluster's region, and then click the cluster ID to open the details page.
In the left-side navigation pane, choose Performance Monitoring > Search Clusters > Synchronization Link.

The following table describes the metrics.

Metric	Unit	Description
Sync Job Restart Count	Count	The number of restarts for the data synchronization link between PolarDB and PolarSearch. Frequent restarts can indicate an unstable link.
Sync Job Latency	ms	The latency of the data synchronization link between PolarDB and PolarSearch.
Sink Output Records per Second	Count/s	The number of records that the data synchronization link writes to PolarSearch per second.
Source Input Records Per Second	Count/s	The number of records that the data synchronization link reads from PolarDB per second.
Completed Checkpoints	Count	The number of checkpoints completed by the data synchronization link. If a job fails, it automatically restarts from the last completed checkpoint.
Is Data Source Incremental	0 or 1	Indicates whether data synchronization is in the incremental phase. A value of `0` indicates the full synchronization phase, and a value of `1` indicates the incremental synchronization phase.

Note

Error messages returned from the kernel may be delayed by one to two minutes.

Configure alert rules

You can use CloudMonitor to configure alert rules for each synchronization link, ensuring timely notification of any issues. You can access the CloudMonitor console from the metrics page in the PolarDB console, or go to it directly.

Log on to the CloudMonitor console.
In Alerts, select Alert Rules, and then click Create Alert Rule.
From the Product drop-down list, select PolarDB - Search.
Select the target cluster ID, and configure the alert metrics and trigger conditions for the synchronization job.
Configure notification methods, such as an alert contact group or webhook callback, and then click Confirm to create the rule.

Note

The metrics available for alert rules are the same as those on the Performance Monitoring page in the PolarDB console. We recommend configuring alerts for Sync Job Restart Count and Sync Job Latency to promptly detect synchronization issues.