Elasticsearch WARN log : "component": "o.e.c.s.ClusterApplierService" "message": "cluster state applier task [Publication{term=17, version=7278}] took [41.2s] which is above the warn threshold of [30s]

[Christian_Dahlqvist]

This seem to indicate that it takes a very long time persisting a cluster state update. This is quite possibly due to slow or overloaded storage.