O&M dashboard - DataWorks - Alibaba Cloud Documentation Center

Usage notes

With the O&M dashboard, you can view the operations of auto triggered nodes, manually triggered nodes, and Data Integration tasks in your workspace from the following perspectives:

Specified workspace: View the O&M overview for a selected workspace. This view provides an O&M summary for the entire workspace and allows you to check the O&M details of Data Integration synchronization tasks separately.
All workspaces: View the O&M overview for all workspaces under your current account. In this view, you cannot separately check the O&M details of Data Integration synchronization tasks.

Limitations

The O&M dashboard is not available in the development environment of a workspace in standard mode.

Note
You can switch between the Production and Development environments in the top menu bar of the Operation Center interface.
Auto Triggered Node tab: Displays O&M data only for auto triggered nodes and their instances. Data from other types of tasks and instances is excluded.
Manually Triggered Node tab: Displays O&M data only for manually triggered workflows and their internal node instances.
Data Integration tab: Displays O&M data only for batch synchronization and real-time synchronization tasks in Data Integration.

Access the O&M dashboard

Log on to the DataWorks console. In the target region, click Data Development and O&M > Operation Center in the left-side navigation pane. Select a workspace from the drop-down list and click Go to Operation Center.

O&M for auto triggered nodes

The Auto Triggered Node tab shows the O&M overview, including the stability assessment, key issues, instance status distribution, instance completion status, and resource usage for resource groups for scheduling.

O&M stability assessment

The system assesses your workspace's operational stability based on the overall run status of its tasks.

Workspace	Single workspace	All My Workspaces
Stability chart		Lists all workspaces with their name, stability assessment level, number of auto triggered instances, auto triggered instance completion rate, and workspace health. The list supports paging. Click View Details on the right of a workspace to open the O&M details of that single workspace.
Description	The stability health is rated as Excellent, Good, Fair, or Poor. High-risk or low-risk labels indicate a poor health status that requires immediate attention and optimization.	Switch to the All My Workspaces view at the top of the page to view the O&M stability, number of auto triggered instances, and instance completion status for all workspaces you have joined. You can also click View Details in the Actions column for a specific workspace to view its O&M stability details.

Key issues

This section highlights operational exceptions for auto triggered nodes from both a workspace-wide and a personal perspective, based on smart baselines and exception statistics. You can view overall issues in the workspace or only the issues related to tasks you own. This allows you to identify and resolve problems promptly to prevent business impact.

Issue type	Description	Related documentation	Illustration
Baseline instance in overtime	The number of baseline instances that are in overtime today. An instance is in overtime if its estimated completion time exceeds the baseline's committed time, which triggers an alert.	Baseline instances	Displays eight key metrics as cards: the number of Baseline instances in overtime, Baseline instances in alert, Error events, and Slowdown events, each with its day-over-day and week-over-week change, plus the number of Isolated nodes, Frozen nodes, Expired nodes, and Nodes modified today. You can switch between the All and Mine views.
Baseline instance in alert	The number of baseline instances in an alert state today. The alert margin helps ensure that important data is produced on time in complex dependency scenarios. If the alert margin is exceeded, nodes may fail to complete on time and exceptions may occur.	Baseline committed time and alert margin
Error event	The number of error events that occurred today. When a task is monitored by a baseline, a run failure generates an error event. A failed task can block its downstream dependencies, so you must resolve it promptly to ensure normal operation.	Event management
Slowdown event	The number of slowdown events that occurred today. A slowdown event occurs when a baseline-monitored task's run duration significantly increases compared to its historical average.	Event management
Isolated node	The number of auto triggered nodes that have no upstream dependencies. An auto triggered node without upstream dependencies becomes an isolated node and will no longer run automatically.	Isolated nodes
Frozen node	The number of auto triggered nodes in the Frozen (paused) state. When an auto triggered node is frozen, its instances are also frozen. Frozen instances do not run and will block their downstream dependencies.	Freeze and unfreeze tasks
Expired node	The number of auto triggered nodes whose scheduling validity period has ended. A node automatically generates and runs auto triggered instances within its scheduling validity period. Outside this period, no instances are generated or scheduled.	None
Modified node	The number of auto triggered nodes modified today. Modification operations: Includes changes to code, scheduling configuration, node status, and node owner. Scope: Includes changes made in DataStudio and deployed to the production environment, as well as direct modifications to auto triggered nodes in the production environment. Note When you switch to the Mine view, this metric counts the number of modified nodes for which you are the owner.	None

Overview of auto triggered instances and nodes

The following table describes the O&M overview of auto triggered instances and nodes.

Metric	Description	Illustration
Instances Status	Scope: Shows the status distribution of auto triggered instances for a specific data timestamp, either for the current workspace or for instances you own. The data reflects the state at the time of the page request. How to view: Click a segment of the pie chart to view the number and percentage of instances in that state. Instance states to watch: Failed: A failed instance may block its downstream dependent tasks. Frozen: A frozen instance will not run and will block its downstream dependencies. Slow-running: A running instance is considered slow-running if its duration exceeds its average duration over the past `10` days by more than `15` minutes. If there are fewer than `4` historical instances, an instance is considered slow-running if its duration exceeds 30 minutes. Note This metric includes only normal tasks and excludes dry-run and frozen tasks.
Instances Completion Status	Scope: Shows the completion status (number of successful or not-run instances) of auto triggered instances in the current workspace between `00:00 and 23:00` on the current day, compared to the previous day and the historical average. Display: The line chart shows the completion trends for today, yesterday, and the historical average. Significant deviations between the lines may indicate an anomaly that requires further investigation. Node type: You can filter the results by a specific node type. Historical Average: This is the average completion status over the last `10` days.
Trend of Nodes and Instances	Scope: Shows the trend in the number of auto triggered nodes and instances in the production environment over a selected data timestamp range. You can view data for up to the last year.
Distribution of Auto Triggered Nodes	Scope: Shows the number and proportion of auto triggered nodes, categorized by dimensions like node type and scheduling cycle. The data reflects the state at the time of the page request. Display: The pie chart has a limit on the number of categories it can display. If the number of types exceeds the limit, they will be grouped together. Note In the All My Workspaces view, you can see the distribution of auto triggered nodes by workspace.

Scheduling resource group usage

This section shows the utilization rate of the selected resource group for scheduling (the percentage of resources used by instances running on the group) and the trend of the number of instances running on the group over a specified time period.

Note

You can view data for up to seven days.
If resource usage exceeds 80%, we recommend scaling out the resource group to prevent resource shortages from affecting task execution.
Resource usage and the number of running instances are measured at the resource group level. For example, if you use an exclusive resource group for scheduling that is shared by multiple workspaces, this chart shows the total resource usage and instance count trend across all those workspaces.

调度资源组使用情况

Auto triggered instance rankings

Ranking of Instances on Previous Day

This section ranks the Top 30 auto triggered instances from the previous day by run duration, resource wait time, and slow-running duration. You can use this ranking to quickly find time-consuming tasks. Click an Instance ID to go to the instance details page and view its run diagnostics.

Note
Slow-running duration: This is the difference between an instance's run duration yesterday and its historical average. Instances are sorted by this difference in descending order.
Ranking of auto triggered instances with the highest error rate in the last month

This section ranks auto triggered instances by error rate over the last month, showing the Top 30 tasks. You can use this to quickly identify tasks with a high error rate, view their details, and find the cause of the errors.

Manually triggered nodes

The Manually Triggered Node tab shows the run status of manually triggered workflows and their internal task instances.

Manual task overview

This section shows the total number of manually triggered workflows and internal task instances that started running from a specified date, along with their success rate.

Run status of workflow instances

Metric

Description

Illustration

Distribution of Workflow Instances by Status

A pie chart shows the distribution of manually triggered workflow instances by run status for a given run date.

Click a segment to navigate to the details page for tasks in that state, where you can view and handle issues. Pay close attention to tasks that have Run failed.
You can view data for up to seven days.
When you switch to the Mine view, this chart shows the status distribution of manually triggered workflow instances for which you are the owner.

Workflow Ranking

Ranks workflows by run duration and failure rate for a specified run date.

You can use this ranking to quickly find slow or high-failure-rate workflows. Click a Node ID to open the Manual Workflow Instance details page. There, you can view the run diagnostics for a specific instance in the workflow's DAG to understand its run status.
Only the Top 30 workflows are displayed.

Lists the manually triggered workflow ranking with fields such as ordinal, workflow name, creator, and run duration. You can switch the sort dimension between Run Duration and Failures, and filter by run date.

Run status of internal task instances

Metric

Description

Illustration

Distribution of Internal Tasks

A pie chart shows the real-time distribution of internal task instances in Operation Center, categorized by Node Type and Owner.

Internal Task Ranking

Ranks internal task instances by run duration and failure rate for a specified run date.

You can use this ranking to quickly find slow or high-failure-rate internal task instances. Click a Node ID to go to the Manual Workflow Instance details page. On the details page, you can view the run diagnostics for a specific instance in the workflow's DAG to understand its run status.
Only the Top 30 internal tasks are displayed.

Lists the internal task ranking with fields such as ordinal, node ID, node name, workflow, node type, creator, and run duration. You can switch the sort dimension between Run Duration and Failures.

Data Integration tasks

The Data Integration tab shows an overview of Data Integration synchronization tasks and resource group usage for Yesterday or Today.

Data Integration resource group usage

This section shows the resource details for all Data Integration tasks in the current workspace, including the Running Tasks, Resource Usage, and Last Modified At. You can use the resource utilization and task volume to determine whether to scale resources up or down for optimal allocation. It also shows information such as the resource group name, status, number of queued tasks, specification, and region, and provides the View, Renew, and Create Resource Group entry points.

Note

For more information about operations on an exclusive resource group for Data Integration, see Billing of exclusive resource groups for Data Integration.
For more information about operations on a serverless resource group, see Use a serverless resource group.
The Dashboard > Data Integration tab displays O&M data only for exclusive resource groups for Data Integration.

Run status distribution of Data Integration synchronization tasks

A pie chart shows the distribution of synchronization tasks in the current workspace by run status. Click a segment to navigate to the details page for tasks in that state to view and handle issues. Pay close attention to Abnormal and Run failed tasks, as they often block downstream task execution. 运行状态分布

Run status of batch synchronization tasks

The following table describes the run status of batch synchronization tasks.

Metric	Description	Illustration
Data Synchronization Progress	Shows the total amount of data and total traffic used for batch synchronization within the selected data timestamp.	Displays three batch synchronization progress metrics as cards: Total Data Volume, Total Public Network Traffic, and Total Records.
Statistics on amount of synchronized data	Displays data read and write curves by data source type for the selected data timestamp. This helps you quickly identify engine tasks with large data volumes and determine if they require more resources.
Latest Top 10 Tasks	Shows the 10 most recent Latest Failed Instance and Latest Successful Instance to give you a global view of the latest synchronization task statuses. You can use the error messages to quickly identify the cause of a failure and resolve it.	Lists the Recent Failed Instances and Recent Successful Instances with fields such as node name, node ID, start time, end time, and error message.
Execution details of synchronization tasks	You can quickly search for task instances by filtering on conditions such as Commission Time, Node Status, and Node Name to view their running details.	Lists the execution details of synchronization tasks with fields such as instance ID, node name, start time, end time, duration, source and destination data sources, synchronized data volume, number of synchronized records, and synchronization rate. You can filter by run date, node status, and node name.

Run status of real-time synchronization tasks

The following table describes the run status of real-time synchronization tasks.

Metric	Description	Illustration
Data synchronization Overview	Shows the total data rate and record rate for all real-time synchronization tasks in the current workspace.	Displays two real-time synchronization overview metrics as cards: Data Speed (BPS) and Record Speed (RPS).
Top 10 tasks with the highest latency	Shows the 10 real-time synchronization tasks with the highest latency, allowing you to quickly identify and optimize them.	Lists the real-time synchronization tasks with the highest latency, including the node name, business latency, and resource group information.
Alerts	Shows recent alert information from real-time synchronization tasks, allowing you to quickly catch and resolve exceptions.	Lists the alert records with fields such as occurrence time, node ID, node name, recipient, alert type (CRITICAL or WARNING), trigger condition, threshold, alert time, and notification method.
Failover Information	Shows `failover` messages for real-time synchronization tasks within a specified time frame to provide an overview of task `failover` status. For more information about `failover`, see Run and manage real-time synchronization tasks.	Lists the `failover` records within the specified time range with fields such as time, instance ID, node name, failover event, and actions.