Enable concurrent_segment_search auto mode by default

Signed-off-by: Vikasht34 <viktari@amazon.com>

Author: Vikasht34

CHANGELOG.md

@@ -11,6 +11,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 ### Dependencies
 
 ### Changed
+* Enable concurrent_segment_search auto mode by default[#17978](https://github.com/opensearch-project/OpenSearch/pull/17978)
 
 ### Deprecated
 

server/src/main/java/org/opensearch/search/SearchService.java

@@ -279,7 +279,7 @@ public class SearchService extends AbstractLifecycleComponent implements IndexEv
 
     public static final Setting<String> CLUSTER_CONCURRENT_SEGMENT_SEARCH_MODE = Setting.simpleString(
         "search.concurrent_segment_search.mode",
-        CONCURRENT_SEGMENT_SEARCH_MODE_NONE,
+        CONCURRENT_SEGMENT_SEARCH_MODE_AUTO,
         value -> {
             switch (value) {
                 case CONCURRENT_SEGMENT_SEARCH_MODE_ALL:
@@ -303,12 +303,11 @@ public class SearchService extends AbstractLifecycleComponent implements IndexEv
     // value == 0 means lucene slice computation will be used
     public static final Setting<Integer> CONCURRENT_SEGMENT_SEARCH_TARGET_MAX_SLICE_COUNT_SETTING = Setting.intSetting(
         CONCURRENT_SEGMENT_SEARCH_TARGET_MAX_SLICE_COUNT_KEY,
-        CONCURRENT_SEGMENT_SEARCH_TARGET_MAX_SLICE_COUNT_DEFAULT_VALUE,
+        computeDefaultSliceCount(),
         CONCURRENT_SEGMENT_SEARCH_TARGET_MAX_SLICE_COUNT_DEFAULT_VALUE,
         Property.Dynamic,
         Property.NodeScope
     );
-
     // value 0 means rewrite filters optimization in aggregations will be disabled
     @ExperimentalApi
     public static final Setting<Integer> MAX_AGGREGATION_REWRITE_FILTERS = Setting.intSetting(
@@ -1871,6 +1870,27 @@ public MinAndMax<?> estimatedMinAndMax() {
         }
     }
 
+    /**
+     * Computes the default maximum number of slices for concurrent segment search.
+     * <p>
+     * This value is dynamically calculated as:
+     * <pre>
+     *     min(availableProcessors / 2, 4)
+     * </pre>
+     * This ensures that:
+     * <ul>
+     *   <li>On small machines, it avoids over-threading.</li>
+     *   <li>On larger machines, it caps the concurrency to a reasonable level (4 slices).</li>
+     * </ul>
+     * This default is used when the user does not explicitly set the
+     * {@code search.concurrent.max_slice_count} cluster setting.
+     *
+     * @return the computed default slice count
+     */
+    private static int computeDefaultSliceCount() {
+        return Math.min(Runtime.getRuntime().availableProcessors() / 2, 4);
+    }
+
     /**
      * This helper class ensures we only execute either the success or the failure path for {@link SearchOperationListener}.
      * This is crucial for some implementations like {@link org.opensearch.index.search.stats.ShardSearchStats}.