Add a metastore interface to control add partition commit batch size

Reference to [HMS Repair Utility](shorturl.at/eHIXY), the size for adding partitions
can be in a range between 1 and 2,147,483,647.
For [Glue metastore](shorturl.at/pqDH9), it’s bounded by 100 for write access.

presto-hive-metastore/src/main/java/com/facebook/presto/hive/metastore/ExtendedHiveMetastore.java

@@ -145,4 +145,11 @@ default List<TableConstraint<String>> getTableConstraints(MetastoreContext metas
     {
         return ImmutableList.of();
     }
+
+    // Different metastore systems could implement this commit batch size differently based on different underlying database capacity.
+    // Default batch partition commit size is set to 10.
+    default int getPartitionCommitBatchSize()
+    {
+        return 10;
+    }
 }

presto-hive-metastore/src/main/java/com/facebook/presto/hive/metastore/SemiTransactionalHiveMetastore.java

@@ -115,7 +115,6 @@
 public class SemiTransactionalHiveMetastore
 {
     private static final Logger log = Logger.get(SemiTransactionalHiveMetastore.class);
-    private static final int PARTITION_COMMIT_BATCH_SIZE = 8;
     private static final int MAX_LAST_DATA_COMMIT_TIME_ENTRY_PER_TABLE = 100;
     private static final int MAX_LAST_DATA_COMMIT_TIME_ENTRY_PER_TRANSACTION = 10_000;
 
@@ -1513,7 +1512,7 @@ private void prepareAddPartition(MetastoreContext metastoreContext, HdfsContext
 
             PartitionAdder partitionAdder = partitionAdders.computeIfAbsent(
                     partition.getSchemaTableName(),
-                    ignored -> new PartitionAdder(metastoreContext, partition.getDatabaseName(), partition.getTableName(), delegate, PARTITION_COMMIT_BATCH_SIZE));
+                    ignored -> new PartitionAdder(metastoreContext, partition.getDatabaseName(), partition.getTableName(), delegate));
 
             // we can bypass the file storage path checking logic for sync partition code path
             // because the file paths have been verified during early phase of the sync logic already
@@ -2961,13 +2960,13 @@ private static class PartitionAdder
         private List<List<String>> createdPartitionValues = new ArrayList<>();
         private List<MetastoreOperationResult> operationResults;
 
-        public PartitionAdder(MetastoreContext metastoreContext, String schemaName, String tableName, ExtendedHiveMetastore metastore, int batchSize)
+        public PartitionAdder(MetastoreContext metastoreContext, String schemaName, String tableName, ExtendedHiveMetastore metastore)
         {
             this.metastoreContext = requireNonNull(metastoreContext, "metastoreContext is null");
             this.schemaName = schemaName;
             this.tableName = tableName;
             this.metastore = metastore;
-            this.batchSize = batchSize;
+            this.batchSize = metastore.getPartitionCommitBatchSize();
             this.partitions = new ArrayList<>(batchSize);
             this.operationResults = new ArrayList<>();
         }

presto-hive-metastore/src/main/java/com/facebook/presto/hive/metastore/glue/GlueHiveMetastore.java

@@ -150,7 +150,24 @@ public class GlueHiveMetastore
     private static final String PUBLIC_ROLE_NAME = "public";
     private static final String DEFAULT_METASTORE_USER = "presto";
     private static final String WILDCARD_EXPRESSION = "";
+
+    // This is the total number of partitions allowed to process in a big batch chunk which splits multiple smaller batch of partitions allowed by BATCH_CREATE_PARTITION_MAX_PAGE_SIZE
+    // Here's an example diagram on how async batches are handled for Create Partition:
+    // |--------BATCH_CREATE_PARTITION_MAX_PAGE_SIZE------------| ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+    // | p0, p1, p2 .....................................   p99 |
+    // |--------------------------------------------------------|
+    // | p0, p1, p2 .....................................   p99 |
+    // |--------------------------------------------------------|
+    // BATCH_PARTITION_COMMIT_TOTAL_SIZE / BATCH_CREATE_PARTITION_MAX_PAGE_SIZE ..... (10k/100=100 batches)
+    // |--------------------------------------------------------|++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+    // | p0, p1, p2 .....................................   p99 |
+    // |--------------------------------------------------------|
+    // | p0, p1, p2 .....................................   p99 |
+    // |--------------------------------------------------------|.......... (100 batches)
+    // |--------------------------------------------------------|++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+    private static final int BATCH_PARTITION_COMMIT_TOTAL_SIZE = 10000;
     private static final int BATCH_GET_PARTITION_MAX_PAGE_SIZE = 1000;
+    // this is the total number of partitions allowed per batch that glue metastore can process to create partitions
     private static final int BATCH_CREATE_PARTITION_MAX_PAGE_SIZE = 100;
     private static final int AWS_GLUE_GET_PARTITIONS_MAX_RESULTS = 1000;
     private static final Comparator<Partition> PARTITION_COMPARATOR = comparing(Partition::getValues, lexicographical(String.CASE_INSENSITIVE_ORDER));
@@ -226,6 +243,14 @@ public GlueMetastoreStats getStats()
         return stats;
     }
 
+    // For Glue metastore there's an upper bound limit on 100 partitions per batch.
+    // Here's the reference: https://docs.aws.amazon.com/glue/latest/webapi/API_BatchCreatePartition.html
+    @Override
+    public int getPartitionCommitBatchSize()
+    {
+        return BATCH_PARTITION_COMMIT_TOTAL_SIZE;
+    }
+
     @Override
     public Optional<Database> getDatabase(MetastoreContext metastoreContext, String databaseName)
     {