I have more than 100 instances that should be terminated. Those instances

If you would like to Complete all the active instances rather than Cancel or Delete, you may perform the following steps

1. In the Main process model, add a "Script Task" before an End event and set End Event to be a Terminate Event. This would ensure that all active instances will be Completed when the script task is triggered. Save and publish a new version.

Note the UUID of the script task which can be found by going into the process model documentation of the main model and clicking on the script task.

2. Download the Process Upgrade Plugin (Process Upgrade_1.0.3.jar) from Appian Forum > Records > Shared Components

3. Copy into <APPIAN_HOME>/_admin/plugins folder

4. Create a process model to perform a Process Upgrade adding an "Upgrade Processes" node and a "Start All Nodes" between the start and end events.

5. In the "Upgrade Processes" Node
- Set Dry Run to false
- Specify the Process ids of active instances
- Specify the main Process Model id

6. In the "Start all Nodes" Node
- Set Dry Run to false
- Specify the UUID of the script task that was added in the main model
- Specify the Process ids of actve instances

7. Set isUpgrade to true and run the upgrade model to update instances to the latest version

8. Set isUpgrade to false and run the upgrade model again to complete the active instances



7. Run the Process Upgrade Process model from, which would upgrade all previous instances to the current version of the process model.
8. Send a message to the process model to initiate Completion of active processes.

Thanks Eduardo, As this is in production environment, we are not allowed to publish the process model. and more over there are other instances that do not require termination. (we have approx. 10k instances for that main process and we need to terminate a few hundreds among them that meet certain conditions.)

The only constraint here is we should terminate the unwanted instances with out publishing the main process model.

The problem is that the "Start All Nodes" node cannot take the UUID of an END EVENT. If you have a node before the end event that can be re-executed you can do still use this approach but basically use the Start All Nodes smart service.