SERVER-72146 Add extra migration statistics including the total bytes cloned before and during catch up, committed and aborted migrations and total time

(cherry picked from commit 9fae6f63e917c33d6b4bee6d57446a3b6111b96a)

5 files changed, 39 insertions, 2 deletions

diff --git a/src/mongo/db/s/migration_batch_inserter.cpp b/src/mongo/db/s/migration_batch_inserter.cpp
index a73d2b52eec..37b2c0d07dd 100644
--- a/src/mongo/db/s/migration_batch_inserter.cpp
+++ b/src/mongo/db/s/migration_batch_inserter.cpp
@@ -165,6 +165,7 @@ void MigrationBatchInserter::run(Status status) const try {
             repl::ReplClientInfo::forClient(opCtx->getClient()).getLastOp());
 
         ShardingStatistics::get(opCtx).countDocsClonedOnRecipient.addAndFetch(batchNumCloned);
+        ShardingStatistics::get(opCtx).countBytesClonedOnRecipient.addAndFetch(batchClonedBytes);
         LOGV2(6718408,
               "Incrementing numCloned count by {batchNumCloned} and numClonedBytes by "
               "{batchClonedBytes}",
diff --git a/src/mongo/db/s/migration_destination_manager.cpp b/src/mongo/db/s/migration_destination_manager.cpp
index 4c384e7b2a4..c075a88c5da 100644
--- a/src/mongo/db/s/migration_destination_manager.cpp
+++ b/src/mongo/db/s/migration_destination_manager.cpp
@@ -1438,6 +1438,8 @@ void MigrationDestinationManager::_migrateDriver(OperationContext* outerOpCtx,
                 if (!_applyMigrateOp(opCtx, nextBatch)) {
                     return true;
                 }
+                ShardingStatistics::get(opCtx).countBytesClonedOnCatchUpOnRecipient.addAndFetch(
+                    nextBatch["size"].number());
 
                 const int maxIterations = 3600 * 50;
 
@@ -1693,6 +1695,7 @@ void MigrationDestinationManager::_migrateDriver(OperationContext* outerOpCtx,
 bool MigrationDestinationManager::_applyMigrateOp(OperationContext* opCtx, const BSONObj& xfer) {
     bool didAnything = false;
     long long changeInOrphans = 0;
+    long long totalDocs = 0;
 
     // Deleted documents
     if (xfer["deleted"].isABSONObj()) {
@@ -1703,6 +1706,7 @@ bool MigrationDestinationManager::_applyMigrateOp(OperationContext* opCtx, const
 
         BSONObjIterator i(xfer["deleted"].Obj());
         while (i.more()) {
+            totalDocs++;
             AutoGetCollection autoColl(opCtx, _nss, MODE_IX);
             uassert(ErrorCodes::ConflictingOperationInProgress,
                     str::stream() << "Collection " << _nss.ns()
@@ -1745,6 +1749,7 @@ bool MigrationDestinationManager::_applyMigrateOp(OperationContext* opCtx, const
     if (xfer["reload"].isABSONObj()) {
         BSONObjIterator i(xfer["reload"].Obj());
         while (i.more()) {
+            totalDocs++;
             AutoGetCollection autoColl(opCtx, _nss, MODE_IX);
             uassert(ErrorCodes::ConflictingOperationInProgress,
                     str::stream() << "Collection " << _nss.ns()
@@ -1792,6 +1797,9 @@ bool MigrationDestinationManager::_applyMigrateOp(OperationContext* opCtx, const
     if (changeInOrphans != 0) {
         persistUpdatedNumOrphans(opCtx, *_collectionUuid, ChunkRange(_min, _max), changeInOrphans);
     }
+
+    ShardingStatistics::get(opCtx).countDocsClonedOnCatchUpOnRecipient.addAndFetch(totalDocs);
+
     return didAnything;
 }
 
diff --git a/src/mongo/db/s/migration_util.cpp b/src/mongo/db/s/migration_util.cpp
index 5f6cd2253e6..1d6690fc571 100644
--- a/src/mongo/db/s/migration_util.cpp
+++ b/src/mongo/db/s/migration_util.cpp
@@ -797,6 +797,7 @@ void persistCommitDecision(OperationContext* opCtx,
         store.update(opCtx,
                      BSON(MigrationCoordinatorDocument::kIdFieldName << migrationDoc.getId()),
                      migrationDoc.toBSON());
+        ShardingStatistics::get(opCtx).countDonorMoveChunkCommitted.addAndFetch(1);
     } catch (const ExceptionFor<ErrorCodes::NoMatchingDocument>&) {
         LOGV2_ERROR(6439800,
                     "No coordination doc found on disk for migration",
@@ -820,6 +821,7 @@ void persistAbortDecision(OperationContext* opCtx,
         store.update(opCtx,
                      BSON(MigrationCoordinatorDocument::kIdFieldName << migrationDoc.getId()),
                      migrationDoc.toBSON());
+        ShardingStatistics::get(opCtx).countDonorMoveChunkAborted.addAndFetch(1);
     } catch (const ExceptionFor<ErrorCodes::NoMatchingDocument>&) {
         LOGV2(6439801,
               "No coordination doc found on disk for migration",
diff --git a/src/mongo/db/s/sharding_statistics.cpp b/src/mongo/db/s/sharding_statistics.cpp
index 5f328c17e4a..f56ea543378 100644
--- a/src/mongo/db/s/sharding_statistics.cpp
+++ b/src/mongo/db/s/sharding_statistics.cpp
@@ -54,6 +54,9 @@ void ShardingStatistics::report(BSONObjBuilder* builder) const {
     builder->append("countStaleConfigErrors", countStaleConfigErrors.load());
 
     builder->append("countDonorMoveChunkStarted", countDonorMoveChunkStarted.load());
+    builder->append("countDonorMoveChunkCommitted", countDonorMoveChunkCommitted.load());
+    builder->append("countDonorMoveChunkAborted", countDonorMoveChunkAborted.load());
+    builder->append("totalDonorMoveChunkTimeMillis", totalDonorMoveChunkTimeMillis.load());
     builder->append("totalDonorChunkCloneTimeMillis", totalDonorChunkCloneTimeMillis.load());
     builder->append("totalCriticalSectionCommitTimeMillis",
                     totalCriticalSectionCommitTimeMillis.load());
@@ -61,6 +64,11 @@ void ShardingStatistics::report(BSONObjBuilder* builder) const {
     builder->append("totalRecipientCriticalSectionTimeMillis",
                     totalRecipientCriticalSectionTimeMillis.load());
     builder->append("countDocsClonedOnRecipient", countDocsClonedOnRecipient.load());
+    builder->append("countBytesClonedOnRecipient", countBytesClonedOnRecipient.load());
+    builder->append("countDocsClonedOnCatchUpOnRecipient",
+                    countDocsClonedOnCatchUpOnRecipient.load());
+    builder->append("countBytesClonedOnCatchUpOnRecipient",
+                    countBytesClonedOnCatchUpOnRecipient.load());
     builder->append("countDocsClonedOnDonor", countDocsClonedOnDonor.load());
     builder->append("countRecipientMoveChunkStarted", countRecipientMoveChunkStarted.load());
     builder->append("countDocsDeletedOnDonor", countDocsDeletedOnDonor.load());
diff --git a/src/mongo/db/s/sharding_statistics.h b/src/mongo/db/s/sharding_statistics.h
index 91a10cc4308..7a7efeb1464 100644
--- a/src/mongo/db/s/sharding_statistics.h
+++ b/src/mongo/db/s/sharding_statistics.h
@@ -49,6 +49,12 @@ struct ShardingStatistics {
     // (whether they succeeded or not).
     AtomicWord<long long> countDonorMoveChunkStarted{0};
 
+    // Cumulative, always-increasing counter of how many chunks this node successfully committed.
+    AtomicWord<long long> countDonorMoveChunkCommitted{0};
+
+    // Cumulative, always-increasing counter of how many move chunks this node aborted.
+    AtomicWord<long long> countDonorMoveChunkAborted{0};
+
     // Cumulative, always-increasing counter of how much time the entire move chunk operation took
     // (excluding range deletion).
     AtomicWord<long long> totalDonorMoveChunkTimeMillis{0};
@@ -61,12 +67,24 @@ struct ShardingStatistics {
     // recipient node.
     AtomicWord<long long> countDocsClonedOnRecipient{0};
 
+    // Cumulative, always-increasing counter of how many documents have been cloned on the catch up
+    // phase on the recipient node.
+    AtomicWord<long long> countDocsClonedOnCatchUpOnRecipient{0};
+
+    // Cumulative, always-increasing counter of how many bytes have been cloned on the catch up
+    // phase on the recipient node.
+    AtomicWord<long long> countBytesClonedOnCatchUpOnRecipient{0};
+
+    // Cumulative, always-increasing counter of how many bytes have been cloned on the
+    // recipient node.
+    AtomicWord<long long> countBytesClonedOnRecipient{0};
+
     // Cumulative, always-increasing counter of how many documents have been cloned on the donor
     // node.
     AtomicWord<long long> countDocsClonedOnDonor{0};
 
-    // Cumulative, always-increasing counter of how many documents have been deleted on the donor
-    // node by the rangeDeleter.
+    // Cumulative, always-increasing counter of how many documents have been deleted by the
+    // rangeDeleter.
     AtomicWord<long long> countDocsDeletedOnDonor{0};
 
     // Cumulative, always-increasing counter of how many chunks this node started to receive