diff --git a/be/src/common/config.h b/be/src/common/config.h
index dbc9887d08..b59d9d42e3 100644
--- a/be/src/common/config.h
+++ b/be/src/common/config.h
@@ -937,6 +937,9 @@ CONF_Int32(max_depth_of_expr_tree, "600");
 // Report a tablet as bad when io errors occurs more than this value.
 CONF_mInt64(max_tablet_io_errors, "-1");
 
+// Page size of row column, default 4KB
+CONF_mInt64(row_column_page_size, "4096");
+
 #ifdef BE_TEST
 // test s3
 CONF_String(test_s3_resource, "resource");
diff --git a/be/src/olap/rowset/segment_v2/segment_writer.cpp b/be/src/olap/rowset/segment_v2/segment_writer.cpp
index 71b2939f18..0789777cf3 100644
--- a/be/src/olap/rowset/segment_v2/segment_writer.cpp
+++ b/be/src/olap/rowset/segment_v2/segment_writer.cpp
@@ -17,6 +17,7 @@
 
 #include "olap/rowset/segment_v2/segment_writer.h"
 
+#include "common/config.h"
 #include "common/consts.h"
 #include "common/logging.h" // LOG
 #include "io/fs/file_writer.h"
@@ -189,7 +190,7 @@ Status SegmentWriter::init(const std::vector<uint32_t>& col_ids, bool has_key,
 
         if (column.is_row_store_column()) {
             // smaller page size for row store column
-            opts.data_page_size = 16 * 1024;
+            opts.data_page_size = config::row_column_page_size;
         }
 
         std::unique_ptr<ColumnWriter> writer;
diff --git a/be/src/olap/tablet.cpp b/be/src/olap/tablet.cpp
index f70bc7f4c7..2bf536468f 100644
--- a/be/src/olap/tablet.cpp
+++ b/be/src/olap/tablet.cpp
@@ -2329,7 +2329,8 @@ TabletSchemaSPtr Tablet::get_max_version_schema(std::lock_guard<std::shared_mute
 
 Status Tablet::lookup_row_data(const Slice& encoded_key, const RowLocation& row_location,
                                RowsetSharedPtr input_rowset, const TupleDescriptor* desc,
-                               vectorized::Block* block, bool write_to_cache) {
+                               OlapReaderStatistics& stats, vectorized::Block* block,
+                               bool write_to_cache) {
     // read row data
     BetaRowsetSharedPtr rowset = std::static_pointer_cast<BetaRowset>(input_rowset);
     if (!rowset) {
@@ -2366,7 +2367,6 @@ Status Tablet::lookup_row_data(const Slice& encoded_key, const RowLocation& row_
                                                      &column_iterator));
         std::unique_ptr<segment_v2::ColumnIterator> ptr_guard(column_iterator);
         segment_v2::ColumnIteratorOptions opt;
-        OlapReaderStatistics stats;
         opt.file_reader = segment->file_reader().get();
         opt.stats = &stats;
         opt.use_page_cache = !config::disable_storage_page_cache;
diff --git a/be/src/olap/tablet.h b/be/src/olap/tablet.h
index 93c7e78192..faa5a5c8e0 100644
--- a/be/src/olap/tablet.h
+++ b/be/src/olap/tablet.h
@@ -382,7 +382,8 @@ public:
     // Lookup a row with TupleDescriptor and fill Block
     Status lookup_row_data(const Slice& encoded_key, const RowLocation& row_location,
                            RowsetSharedPtr rowset, const TupleDescriptor* desc,
-                           vectorized::Block* block, bool write_to_cache = false);
+                           OlapReaderStatistics& stats, vectorized::Block* block,
+                           bool write_to_cache = false);
 
     // calc delete bitmap when flush memtable, use a fake version to calc
     // For example, cur max version is 5, and we use version 6 to calc but
diff --git a/be/src/service/point_query_executor.cpp b/be/src/service/point_query_executor.cpp
index 1db2881007..f57547d003 100644
--- a/be/src/service/point_query_executor.cpp
+++ b/be/src/service/point_query_executor.cpp
@@ -183,6 +183,7 @@ std::string PointQueryExecutor::print_profile() {
     auto lookup_data_us = _profile_metrics.lookup_data_ns.value() / 1000;
     auto output_data_us = _profile_metrics.output_data_ns.value() / 1000;
     auto total_us = init_us + lookup_key_us + lookup_data_us + output_data_us;
+    auto read_stats = _profile_metrics.read_stats;
     return fmt::format(
             ""
             "[lookup profile:{}us] init:{}us, init_key:{}us,"
@@ -192,10 +193,15 @@ std::string PointQueryExecutor::print_profile() {
             ""
             ""
             ", is_binary_row:{}, output_columns:{}, total_keys:{}, row_cache_hits:{}"
+            ", hit_cached_pages:{}, total_pages_read:{}, compressed_bytes_read:{}, "
+            "io_latency:{}ns, "
+            "uncompressed_bytes_read:{}"
             "",
             total_us, init_us, init_key_us, lookup_key_us, lookup_data_us, output_data_us,
             _hit_lookup_cache, _binary_row_format, _reusable->output_exprs().size(),
-            _row_read_ctxs.size(), _row_cache_hits);
+            _row_read_ctxs.size(), _row_cache_hits, read_stats.cached_pages_num,
+            read_stats.total_pages_num, read_stats.compressed_bytes_read, read_stats.io_ns,
+            read_stats.uncompressed_bytes_read);
 }
 
 Status PointQueryExecutor::_init_keys(const PTabletKeyLookupRequest* request) {
@@ -271,7 +277,8 @@ Status PointQueryExecutor::_lookup_row_data() {
         }
         RETURN_IF_ERROR(_tablet->lookup_row_data(
                 _row_read_ctxs[i]._primary_key, _row_read_ctxs[i]._row_location.value(),
-                *(_row_read_ctxs[i]._rowset_ptr), _reusable->tuple_desc(), _result_block.get(),
+                *(_row_read_ctxs[i]._rowset_ptr), _reusable->tuple_desc(),
+                _profile_metrics.read_stats, _result_block.get(),
                 !config::disable_storage_row_cache /*whether write row cache*/));
     }
     return Status::OK();
diff --git a/be/src/service/point_query_executor.h b/be/src/service/point_query_executor.h
index 9dd5a36822..3f4a0426a1 100644
--- a/be/src/service/point_query_executor.h
+++ b/be/src/service/point_query_executor.h
@@ -23,6 +23,7 @@
 #include "common/status.h"
 #include "gen_cpp/internal_service.pb.h"
 #include "gutil/int128.h"
+#include "olap/olap_common.h"
 #include "olap/rowset/rowset.h"
 #include "olap/tablet.h"
 #include "util/runtime_profile.h"
@@ -234,6 +235,7 @@ struct Metrics {
     RuntimeProfile::Counter lookup_key_ns;
     RuntimeProfile::Counter lookup_data_ns;
     RuntimeProfile::Counter output_data_ns;
+    OlapReaderStatistics read_stats;
 };
 
 // An util to do tablet lookup