Functions and features

Updated at:
Copy as MD

Task orchestration in Data Management (DMS) lets you build, schedule, and monitor multi-step data workflows through a visual interface. With task orchestration, you can:

  • Design task flows on a directed acyclic graph (DAG) by dragging and connecting task nodes

  • Run data migration, data synchronization, and data manipulation tasks within the same flow

  • Schedule flows on a timer or trigger them from external events

  • Define upstream and downstream dependencies to enforce execution order

  • Monitor task status in real time and receive alerts when exceptions occur

  • Perform O&M operations — pausing, terminating, rerunning, freezing, unfreezing, and backfilling data — on running or completed flows

Important

The task orchestration feature is currently in public preview and is free of charge during the preview period.

Key concepts

Concept

Description

Task flow

A complete workflow made up of task nodes and the dependencies between them.

Task node

A single unit of work within a task flow, such as a data migration job or a data quality check.

Directed acyclic graph (DAG)

The visual canvas where you arrange task nodes and draw dependency edges.

Timed scheduling

A scheduling mode that triggers a task flow at fixed intervals or on a set schedule.

Event scheduling

A scheduling mode that triggers a task flow when a specified external event occurs.

Features

Visual flow design

Drag task nodes onto the DAG canvas and connect them to define the execution order. Edit each node's configuration directly in the graph. You can monitor task status in real time from the same view when the flow is running.

Task types

Task orchestration supports three task types to cover common data processing scenarios:

Task type

Description

Data migration

Move data between databases or storage systems.

Data synchronization

Keep data in sync across sources and targets on an ongoing basis.

Data manipulation

Transform and process data within the database.

Scheduling

Task orchestration supports two scheduling modes:

Scheduling mode

Description

Timed scheduling

Triggers a task flow at fixed intervals or on a set schedule. Use this mode for recurring, time-based processing.

Event scheduling

Triggers a task flow when a specified external event occurs. Use this mode to start flows in response to business events.

Task dependencies

Set upstream and downstream dependencies between task nodes to control execution order and ensure tasks run as expected.

Task parameters

Configure each task node with a data source and destination table. Set additional parameters to match your business requirements.

Data quality

Create data quality rules and exception alert policies on task flows to detect and handle data quality issues at the earliest opportunity.

Monitoring and alerting

Track the status of every task node in real time from the task flow monitoring view. Configure alert notifications to be sent when a node fails or encounters an exception.

Supported notification channels: email, SMS, and DingTalk.

Database engine support and service integrations

Task orchestration supports MySQL, Oracle, PostgreSQL, and SQL Server. It integrates with Data Transmission Service (DTS) and Data Disaster Recovery, enabling cross-database workflows that combine multiple Alibaba Cloud services in a single task flow using service portfolios.

O&M operations

The following operations are available on task flows:

Operation

Description

Pause

Suspend the current run.

Terminate

Stop the current run.

Rerun

Run the task flow again.

Freeze

Prevent the task flow from being triggered by its schedule.

Unfreeze

Re-enable a frozen task flow so it resumes scheduled execution.

Backfill data

Run the task flow for a specified past time range.