How to Handle System Failures Caused by Hardware Issues in NATS Clusters

[lightwinglc]

I would like to ask how to handle the following issue:
When the host running the NATS server and application experiences a hardware failure (monitoring shows CPU at 100%), we cannot log into the host or proactively kill the application via monitoring processes. Logs from other healthy NATS servers indicate that the faulty NATS server was removed from the cluster due to a "slow consumer." However, external service consumers still report system slowness until the host completely crashes after 10+ minutes, after which they report normal operation.

​Question:
Is there an API interface that can remove both the NATS service and its client connections associated with a specific IP from the cluster?