Use default thread pools for Fleet Integration Knowledge system index

The Fleet Integration Knowledge index is a system index managed by Elasticsearch,
but functionally it behaves like a regular search index. To align with
this behavior, it should use the shared thread pools rather than the
dedicated system thread pool. This ensures consistency with other search
and ingest indices, while keeping the only distinction that it is managed
internally by Elasticsearch.

Author: jimczi

server/src/main/java/org/elasticsearch/indices/ExecutorNames.java

@@ -17,6 +17,15 @@
  */
 public record ExecutorNames(String threadPoolForGet, String threadPoolForSearch, String threadPoolForWrite) {
 
+    /**
+     * The thread pools for typical indices and data streams.
+     */
+    public static final ExecutorNames DEFAULT_INDEX_THREAD_POOLS = new ExecutorNames(
+        ThreadPool.Names.GET,
+        ThreadPool.Names.SEARCH,
+        ThreadPool.Names.WRITE
+    );
+
     /**
      * The thread pools for a typical system index.
      */

x-pack/plugin/fleet/src/main/java/org/elasticsearch/xpack/fleet/Fleet.java

@@ -280,6 +280,9 @@ private static SystemIndexDescriptor fleetIntegrationKnowledgeSystemIndexDescrip
             .setType(Type.EXTERNAL_MANAGED)
             .setAllowedElasticProductOrigins(ALLOWED_PRODUCTS)
             .setOrigin(FLEET_ORIGIN)
+            // This is a regular search index so it uses the shared thread pools.
+            // The only difference is that its mappings and settings are managed internally by Elasticsearch.
+            .setThreadPools(ExecutorNames.DEFAULT_INDEX_THREAD_POOLS)
             .setMappings(request.mappings())
             .setSettings(request.settings())
             .setPrimaryIndex(".integration_knowledge-" + CURRENT_INDEX_VERSION)