We have set our agents to terminate after being idle for 30 minutes under the EC2 plugin settings page.

We also have "Free Space Threshold" enabled under "Preventive Node Monitoring", to automatically mark agents as offline whose disks fall under the threshold.

We treat our agents as cattle, so ideally agents that are marked as offline would eventually terminate after being idle for 30 minutes.

But that doesn't seem to be the case when both features are enabled. It seems that agents that are marked offline are being ignored by the idle termination check and are left running indefinitely.