Core: fix SerialMergeScheduler to trigger index throttling

SerialMergeScheduler (deprecated but still accepted, and randomly
picked by tests) was failing to cause index throttling when merges
were falling behind.  This could then cause "Delete Index failed - not
acked" timeouts when it took more than 30 seconds to close the index.

Closes elastic#8405

src/main/java/org/elasticsearch/index/merge/scheduler/SerialMergeSchedulerProvider.java

@@ -56,7 +56,7 @@ public SerialMergeSchedulerProvider(ShardId shardId, @IndexSettings Settings ind
 
     @Override
     public int getMaxMerges() {
-        return 2;
+        return 1;
     }
 
     @Override
@@ -90,7 +90,7 @@ public void close() {
 
     }
 
-    /** NOTE: subclasses TrackingCONCURRENTMergeScheduler, but we set max_merge_count = max_thread_count = 1 above */
+    /** NOTE: subclasses TrackingCONCURRENTMergeScheduler, but we set max_merge_count = 2 and max_thread_count = 1, above */
     public static class CustomSerialMergeScheduler extends TrackingConcurrentMergeScheduler {
 
         private final ShardId shardId;