Task management

Task scheduling mode

• Trigger method: Currently, you can create tasks for scheduling a service robot only by using OpenAPI (see CreateTask) or the MCP Tool (see Publish and manage MCP Tool).

• Task assignment mode: When creating a task, you can either specify a robot manually or have the system assign one automatically.

  Assignment method	Specify robot	Auto-assign by group
  Use cases	The automated process involves software or systems with strict login state restrictions. For example, logging in to an instant messaging (IM) application on different computers can trigger application-specific limitations.	The automated process involves software or systems with no login state restrictions, such as collecting publicly available information from the internet.
  Using OpenAPI	When you call the API, set AllocateMode=manu and specify a RobotId. If the specified robot is idle, the task starts immediately. If the robot is running, the system places the task in its waiting queue. You can obtain the RobotId from the service robot list page in the console or by calling an API.	When you call the API, set AllocateMode=auto. The system then automatically finds a suitable robot in the specified robot group and assigns the task. You can divide service robots into different robot groups. Each group assigns tasks based on its own waiting queue. The system finds an idle robot to execute a task based on task priority. If no robots are idle, the task remains in a pending state.
  Using MCP Tool	When using the MCP Tool, you must specify a robot ID.	Not supported
  Pending task limit	Each robot can have a maximum of 100 pending tasks. If this limit is exceeded, you cannot create new tasks.	Each robot group can have a maximum of 1,000 pending tasks. If this limit is exceeded, you cannot create new tasks.

• Task execution order

  • Basic method: Control the execution order of tasks in the same waiting queue by setting task priority.

    • This applies to tasks within the same waiting queue, such as a specific robot's pending task queue or a robot group's pending task queue.

    • You can set the priority to an integer from 1 to 100. The system sorts tasks based on two rules: first by priority in descending order (a higher number indicates a higher priority), and then by creation time for tasks with the same priority (first in, first out).

    • If both a specific robot and its group have pending tasks, the system executes the tasks assigned directly to the robot first.

  • Advanced method: Use robot groups for enhanced task control.

    • For example, you can divide robots into two groups: a high-priority group with a few robots to ensure the immediate execution of critical tasks, and a standard group with many robots that processes tasks based on their assigned priority.

Task status

A task can have one of the following four statuses:

• Pending: The task is awaiting robot assignment or is in a specific robot's waiting queue.

• Running: A robot is currently executing the task.

• Succeeded: The task has finished execution. This means the RPA application ran to completion and the robot client did not throw an exception during the process. However, a succeeded status does not guarantee that the RPA application achieved its intended business outcome.

• Failed: The system terminated the task due to an engine error during execution. Possible causes include an application error or manual termination.

Manage tasks

In addition to creating tasks, you can perform the following operations:

• Rerun: Creates a new task by copying the configuration of the original one. This operation does not affect the original task's data. You can only rerun tasks in a terminal state (succeeded or failed).

  Note

  When you rerun a task, all its parameters remain unchanged, including the RPA application version. For example, if the original task used version 0.1.0 of an RPA application and you release version 0.1.1 before the rerun, the new task still uses version 0.1.0.

• Delete: Once a task is deleted, it is permanently removed and cannot be recovered. You cannot delete a running task.

• Stop: When you stop a task, its status changes to failed, and the RPA server instructs the service robot to terminate the current execution.

• Modify priority: You can only modify the priority of pending tasks.

Task result and runtime log

Task result

• Use cases

  • Use this method when tasks are not time-sensitive or require recording only a small amount of information.

  • For time-sensitive scenarios or those that require logging a large amount of data, we recommend reporting results by calling your business APIs directly from the automated process.

• Usage: You must add the necessary code to your automated process. For usage instructions and limitations, see task_result.

Runtime log

• Usage: You must add the necessary code to your automated process. For usage instructions and limitations, see Logger.

• For more information about robot logs, see Robot logs.

Task callback

For more information, see Task.