---
title: Trigger Manager
---
:::important Enterprise Feature
The Trigger Manager application is available under the ClearML Enterprise plan
:::
ClearML's Trigger Manager Application lets you define tasks to be run when predefined events occur. The Trigger Manager
is useful for automating your workflows: trigger training a model whenever a new dataset is tagged with a specific tag,
or running a pipeline when a new model is published.
Each trigger is configured to run a ClearML task when its specifications are met.
The events which can activate a trigger include:
* Status change of an object (task, dataset, or model)
* Tagging of an object
* For task triggers, crossing a specified metric threshold
The app monitors your workspace for trigger events and will launch copies of the specified task when the trigger is activated.
## Trigger Manager Instance Configuration
* **Name** - Name for the app instance. This will appear in the instance list
* **Triggers** - Events whose occurrence will cause a task to be executed
* Task ID - ID of a ClearML task to clone and enqueue
for execution when the trigger is activated.
* Target Project - The project where the task will be cloned to
* Queue - The ClearML Queue to which cloned tasks are enqueued (make sure an agent is assigned to that queue)
* Override Task Hyperparameters - Override parameters in the cloned task. Input the original task's configuration
parameter name (including section name e.g. Args/lr). Use UNIX shell-like syntax (splits on whitespace)
* Trigger Type - Triggers can be activated by model, dataset, and/or task activity. Choose which object type to monitor.
* Object selection criteria:
* Match Project - Monitor objects in projects that match this name only
* Match Name - Monitor objects that match this name. Supports both string and Python regex match
* Trigger on Any Tags - Trigger if the object is tagged with *ANY* of the specified tags (comma separated list)
* Trigger on All Tags - Trigger if the object is tagged with *ALL* of the specified tags (comma separated list)
* Trigger events for Datasets and Models:
* Trigger on Publish - Activate trigger when a Dataset/Model is published
* Trigger on Archive - Activate trigger when a Dataset/Model is archived
* Trigger events for Tasks:
* Trigger on Status Change - Activate trigger when a task's status changes to the selected state.
* Trigger on scalar - Activate trigger when a task's specific metric crosses a threshold:
* Metric - Title of metric
* Variant - Metric's variant (series)
* Condition - Activate trigger if the value goes over/under the specified threshold
* Threshold - The metric threshold to monitor
* Add item - Add another trigger
* **Polling frequency** - Time period in minutes at which the workspace is polled for trigger events
## Dashboard
The Trigger Manager app instance's dashboard displays its console log. The log shows the instance’s activity: periodic
polling, and events triggered
