Action On Task
Action on Task performs START, STOP, or RESTART on a selected non-action task and can either affect only the current runtime state or persist the task autostart state for future agent restarts.
Control task lifecycle from rules and buttons
Use this action when Event Manager rules, schedules, manual controls, or action groups need to start, stop, or restart a detector, recorder, grabber, integration task, or other non-action task. It is intended for lifecycle control, not for changing task configuration values.
Before using it in automation, decide whether the action should change the saved autostart setting. That choice determines whether the task state is only temporary for the current session or persists after an agent restart.
Select the controlled task
Set Target task to the non-action task that should be started, stopped, or restarted. The list intentionally excludes action tasks to prevent recursive action calls.
Choose persistent or runtime control
Enable Affect autostart state for operator-level enable/disable switches. Disable it for temporary runtime pauses, runtime starts, and restart workflows that should not change stored configuration.
Tune restart delay
For Restart, increase Restart delay (ms) when the task needs time to release camera handles, files, sockets, GPU resources, or native libraries before starting again.
Configuration parameters
| Parameter | Required | Description | Default |
|---|---|---|---|
Target task | Yes | Non-action task that will receive the lifecycle command. | None |
Affect autostart state | Yes | When enabled, START and STOP update the saved autostart state and persist the target task configuration. RESTART does not change the saved autostart value. | Enabled |
Ignore auto-start state | Conditional | Visible when Affect autostart state is disabled. For runtime START, allows the task to start even if its saved autostart flag is disabled. | Disabled |
Action | Yes | Lifecycle command to perform: Start, Stop, or Restart. | None |
Restart delay (ms) | Conditional | Delay before starting the task again during runtime restart. Visible when Action is Restart. Use it to avoid immediate reconnect loops and resource conflicts. | 3000 |
Choose between persistent and runtime lifecycle control
With Affect autostart state enabled, START and STOP use engine-level operations that persist the changed autostart flag. This makes the action behave like an operator switch that changes whether the task should come back after the agent restarts.
With Affect autostart state disabled, START, STOP, and RESTART operate on the current runtime state and do not persist the target task configuration. This is better for temporary maintenance, demand-driven processing, and watchdog restarts.
Runtime control
Start, stop, or restart a task for the current agent session without changing stored configuration.
Autostart control
Use START or STOP as a persistent enable/disable switch by changing the saved autostart flag.
Recovery restart
Restart stuck or degraded tasks after errors, reconnects, watchdog events, or operator requests.
Task lifecycle scenarios for automation and operations
Autostart state control
Persistent operator switch
Keep Affect autostart state enabled when the action is used as an operator-level enable or disable switch. Start starts the task and persists autostart on; Stop stops the task and persists autostart off.
Temporary runtime pause
Disable Affect autostart state and use Stop when a rule should pause a task only for the current agent session. If the target task still has autostart enabled, it may start again after an agent restart.
Temporary runtime start
Disable Affect autostart state, set Action to Start, and enable Ignore auto-start state when a normally disabled task should run on demand without changing its saved autostart setting.
Respect configured autostart
Disable both Affect autostart state and Ignore auto-start state when the action should try to start the target task only if its own autostart flag already allows it.
Recovery and scheduling
Recovery restart
Use Restart to stop and start a stuck grabber, detector, recorder, or integration task after an error, timeout, watchdog event, or operator request.
Scheduled resource control
Combine this action with schedules or event rules to start heavy processing only during working hours, stop optional detectors at night, or restart a stream at a known low-traffic time.
Dependent pipeline control
Stop or restart downstream tasks when an upstream source changes state, for example pause recording when a camera source is unavailable or restart image processing after a media source reconnects.
Manual maintenance
Create a manually triggered action for tasks that occasionally need a controlled restart without giving operators direct access to the host OS or service console.
Operational notes
Controls non-action tasks
Target task is selected from non-action tasks. This action is intended to control tasks, not to recursively trigger other actions.
Persistent mode changes autostart
With Affect autostart state enabled, START and STOP persist the changed autostart flag. RESTART restarts the target task but does not change its saved autostart value.
Runtime mode does not persist
With Affect autostart state disabled, START, STOP, and RESTART operate on the current runtime state and do not persist the target task configuration.
Ignore auto-start only affects runtime start
Ignore auto-start state matters only for runtime START. If disabled, a task with autostart off can refuse to start; if enabled, the action starts it anyway for the current session.
Dependencies must be healthy
The target task must still be startable. Its source component and parent task chain must be in a compatible state, otherwise a start request can be ignored or fail.
Use restart delay for resource release
Restart delay (ms) helps avoid immediate reconnect loops or resource conflicts after stopping a camera, process, file writer, or network client.
Terminal in task pipelines
When used inside a task pipeline, the action stops downstream processing after controlling the target task. Place it as a terminal control action unless stopping downstream processing is intentional.
Avoid restart loops
Avoid rules where a task restarts itself or rapidly restarts a parent/source task on every event. Add cooldowns, schedules, or Event Manager conditions around restart rules.
Deleted targets require reconfiguration
If the selected task is deleted or unavailable, initialization fails and the action must be reconfigured with a valid target.