Kill feature for containers spun up by the Google Batch Task Runner (GcpBatchTaskRunner)

[wrussell1999]

Feature description

Last night, I was trying to kill/stop this Kestra Flow, and it's subflows. So I select 'Kill Parent and Subflows', and it said 'Queued up action', but it kept running the GCP Batch Jobs (compute VMs), I had to stop by hand.
I didn't see any logs in GCP regarding Stop Operation or anything. After I did stop by hand, it marked the Subflows as "killed"
Am I using the "kill/stop" functionality correctly? I did wait 5 minutes for it before intervening and I am going to try to test it again/later, but thought I would ask in case this was known problem. Ty 🙏

Currently, Containers spun up by a task runner will run till completion even if the Kestra server would fail or when execution would be killed.

[anna-geller]

The priority is to tackle this for the GcpBatchTaskRunner.

Added separate issues for other providers to make sure we kill the infrastructure resources to avoid unnecessary costs for the user:

• Ensure that AWS Batch infrastructure resources are terminated when the Execution is killed by the user plugin-aws#442
• Ensure that Azure Batch infrastructure resources are terminated when the Execution is killed by the user plugin-azure#93
• Ensure that Cloud Run infrastructure resources are terminated when the Execution is killed by the user plugin-gcp#381
• KubernetesTaskRunner — here I believe we delete the job when execution is killed but let's cross-check