Troubles with heap, gc. Node fault

GET /

{
  "name" : "h1-es01",
  "cluster_name" : "h1",
  "cluster_uuid" : "s71YMgBeQhyRUyYVIzX2sg",
  "version" : {
    "number" : "7.9.1",
    "build_flavor" : "oss",
    "build_type" : "rpm",
    "build_hash" : "083627f112ba94dffc1232e8b42b73492789ef91",
    "build_date" : "2020-09-01T21:22:21.964974Z",
    "build_snapshot" : false,
    "lucene_version" : "8.6.2",
    "minimum_wire_compatibility_version" : "6.8.0",
    "minimum_index_compatibility_version" : "6.0.0-beta1"
  },
  "tagline" : "You Know, for Search"
}

GET /_cat/nodes?v

ip             heap.percent ram.percent cpu load_1m load_5m load_15m node.role master name
h1-es02ip           75          93   6    9.98   10.19     9.93 dimr      -      h1-es02
h1-es03ip           50          83   5   14.42   10.10     9.55 dimr      -      h1-es03
h1-es01ip           74          98   4    4.47    5.53     5.61 dimr      *      h1-es01

GET /_cat/health?v

epoch      timestamp cluster status node.total node.data shards  pri relo init unassign pending_tasks max_task_wait_time active_shards_percent
1608884974 08:29:34  h1      green           3         3   2160 1592    0    0        0             0                  -                100.0%

health green now, but in time of node fault it's red (some of our indexes don't have replicas)

GET /_cat/indices?v