Disk indicator refactoring (#90366) (#90456)

Introduced the DiskHealthAnalyzer which determines the health state, the symptom, the impacts, the diagnoses and the details. We also changed the retrieval of the indices per node to use the RoutingNodes instead of the RoutingTable directly for efficiency.

Co-authored-by: Andrei Dan <andrei.dan@elastic.co>

server/src/main/java/org/elasticsearch/cluster/routing/RoutingNode.java

@@ -371,6 +371,10 @@ public ShardRouting[] copyShards() {
         return shards.values().toArray(EMPTY_SHARD_ROUTING_ARRAY);
     }
 
+    public Index[] copyIndices() {
+        return shardsByIndex.keySet().toArray(Index.EMPTY_ARRAY);
+    }
+
     public boolean isEmpty() {
         return shards.isEmpty();
     }