Why "kubectl exec" command sometimes stuck?

Question

I'm using command kubectl exec --ti pod_name -c container_name -- curl -X GET url to call API of another Pod for ad hoc usage. Sometimes the process seems stuck, any Enter or Ctrl + C have no response. I have to kill the PID to stop it.
Searched for reasons and solutions but still confused. Someone said maybe TCP memory is not enough for too many Kubernetes API calls.
Is this usual? Is it a kubectl issue, or can any setting be applied to upgrade the connection pool (I guess) of AKS API server? Thanks in advance.