Hello,
Did you restarted your Elasticsearch then it was resolved?
I tracked myself with all nodes log files. It looked my warm nodes used all heap memory which i set. L related logs i found were shown as below link.
I guess .monitoring indices were not able to be updated from warm nodes after that happened, then my kibana was ahown above error. Weird thing was kibana displayed Status:RED after that even though _cluster/health execution result via each actual nodes were still Status:green.
I decided to restart warm nodes, then issue was gone.
Thank you for sharing update btw.