SERVER-40986 Include relevant server-wide two phase commit metrics on mongod for serverStatus

7 files changed, 196 insertions, 0 deletions

diff --git a/buildscripts/resmokeconfig/suites/sharding_last_stable_mongos_and_mixed_shards.yml b/buildscripts/resmokeconfig/suites/sharding_last_stable_mongos_and_mixed_shards.yml
index 10f2b52c83a..e0dd0edab39 100644
--- a/buildscripts/resmokeconfig/suites/sharding_last_stable_mongos_and_mixed_shards.yml
+++ b/buildscripts/resmokeconfig/suites/sharding_last_stable_mongos_and_mixed_shards.yml
@@ -118,6 +118,7 @@ selector:
   - jstests/sharding/txn_two_phase_commit_coordinator_shutdown_and_restart.js
   - jstests/sharding/txn_two_phase_commit_failover.js
   - jstests/sharding/txn_two_phase_commit_killop.js
+  - jstests/sharding/txn_two_phase_commit_server_status.js
   - jstests/sharding/txn_recover_decision_using_recovery_router.js
   - jstests/sharding/txn_two_phase_commit_wait_for_majority_commit_after_stepup.js
   - jstests/sharding/txn_commit_optimizations_for_read_only_shards.js
diff --git a/jstests/sharding/txn_two_phase_commit_server_status.js b/jstests/sharding/txn_two_phase_commit_server_status.js
new file mode 100644
index 00000000000..7af87314fc1
--- /dev/null
+++ b/jstests/sharding/txn_two_phase_commit_server_status.js
@@ -0,0 +1,19 @@
+// Basic test that the two-phase commit coordinator metrics fields appear in serverStatus output.
+(function() {
+    "use strict";
+
+    const st = new ShardingTest({shards: 1, config: 1});
+
+    const res = assert.commandWorked(st.shard0.adminCommand({serverStatus: 1}));
+    assert.neq(null, res.twoPhaseCommitCoordinator);
+    assert.eq(0, res.twoPhaseCommitCoordinator.totalCreated);
+    assert.eq(0, res.twoPhaseCommitCoordinator.totalStartedTwoPhaseCommit);
+    assert.neq(null, res.twoPhaseCommitCoordinator.currentInSteps);
+    assert.eq(0, res.twoPhaseCommitCoordinator.currentInSteps.writingParticipantList);
+    assert.eq(0, res.twoPhaseCommitCoordinator.currentInSteps.waitingForVotes);
+    assert.eq(0, res.twoPhaseCommitCoordinator.currentInSteps.writingDecision);
+    assert.eq(0, res.twoPhaseCommitCoordinator.currentInSteps.waitingForDecisionAcks);
+    assert.eq(0, res.twoPhaseCommitCoordinator.currentInSteps.deletingCoordinatorDoc);
+
+    st.stop();
+})();
diff --git a/src/mongo/db/s/SConscript b/src/mongo/db/s/SConscript
index be69b3fbcac..6d89971d298 100644
--- a/src/mongo/db/s/SConscript
+++ b/src/mongo/db/s/SConscript
@@ -132,9 +132,11 @@ env.Library(
         'transaction_coordinator_util.cpp',
         'transaction_coordinator.cpp',
         env.Idlc('transaction_coordinator_document.idl')[0],
+        env.Idlc('transaction_coordinators_stats.idl')[0],
     ],
     LIBDEPS_PRIVATE=[
         '$BUILD_DIR/mongo/db/transaction',
+        '$BUILD_DIR/mongo/db/commands/server_status',
         '$BUILD_DIR/mongo/db/commands/txn_cmd_request',
         '$BUILD_DIR/mongo/db/dbdirectclient',
         '$BUILD_DIR/mongo/db/rw_concern_d',
@@ -428,6 +430,7 @@ env.CppUnitTest(
     ],
     LIBDEPS=[
         '$BUILD_DIR/mongo/db/auth/authmocks',
+        '$BUILD_DIR/mongo/db/commands/server_status',
         '$BUILD_DIR/mongo/s/catalog/sharding_catalog_client_mock',
         '$BUILD_DIR/mongo/s/shard_server_test_fixture',
         'transaction_coordinator',
diff --git a/src/mongo/db/s/server_transaction_coordinators_metrics.cpp b/src/mongo/db/s/server_transaction_coordinators_metrics.cpp
index 5bac2c5c211..ef769c441ee 100644
--- a/src/mongo/db/s/server_transaction_coordinators_metrics.cpp
+++ b/src/mongo/db/s/server_transaction_coordinators_metrics.cpp
@@ -35,6 +35,9 @@
 #include "mongo/db/service_context.h"
 
 namespace mongo {
+
+TransactionCoordinatorsSSS transactionCoordinatorsSSS;
+
 namespace {
 const auto ServerTransactionCoordinatorsMetricsDecoration =
     ServiceContext::declareDecoration<ServerTransactionCoordinatorsMetrics>();
@@ -114,4 +117,24 @@ void ServerTransactionCoordinatorsMetrics::decrementCurrentDeletingCoordinatorDo
     _totalDeletingCoordinatorDoc.fetchAndSubtract(1);
 }
 
+void ServerTransactionCoordinatorsMetrics::updateStats(TransactionCoordinatorsStats* stats) {
+    stats->setTotalCreated(_totalCreated.load());
+    stats->setTotalStartedTwoPhaseCommit(_totalStartedTwoPhaseCommit.load());
+
+    CurrentInSteps currentInSteps;
+    currentInSteps.setWritingParticipantList(_totalWritingParticipantList.load());
+    currentInSteps.setWaitingForVotes(_totalWaitingForVotes.load());
+    currentInSteps.setWritingDecision(_totalWritingDecision.load());
+    currentInSteps.setWaitingForDecisionAcks(_totalWaitingForDecisionAcks.load());
+    currentInSteps.setDeletingCoordinatorDoc(_totalDeletingCoordinatorDoc.load());
+    stats->setCurrentInSteps(currentInSteps);
+}
+
+BSONObj TransactionCoordinatorsSSS::generateSection(OperationContext* opCtx,
+                                                    const BSONElement& configElement) const {
+    TransactionCoordinatorsStats stats;
+    ServerTransactionCoordinatorsMetrics::get(opCtx)->updateStats(&stats);
+    return stats.toBSON();
+}
+
 }  // namespace mongo
diff --git a/src/mongo/db/s/server_transaction_coordinators_metrics.h b/src/mongo/db/s/server_transaction_coordinators_metrics.h
index 9218e5584b4..4fcff15b785 100644
--- a/src/mongo/db/s/server_transaction_coordinators_metrics.h
+++ b/src/mongo/db/s/server_transaction_coordinators_metrics.h
@@ -29,7 +29,9 @@
 
 #pragma once
 
+#include "mongo/db/commands/server_status.h"
 #include "mongo/db/operation_context.h"
+#include "mongo/db/s/transaction_coordinators_stats_gen.h"
 #include "mongo/db/service_context.h"
 
 namespace mongo {
@@ -74,6 +76,11 @@ public:
     void incrementCurrentDeletingCoordinatorDoc();
     void decrementCurrentDeletingCoordinatorDoc();
 
+    /**
+     * Appends the accumulated stats to a transaction coordinators stats object for reporting.
+     */
+    void updateStats(TransactionCoordinatorsStats* stats);
+
 private:
     // The total number of transaction coordinators created on this process since the process's
     // inception.
@@ -99,4 +106,16 @@ private:
     AtomicWord<std::int64_t> _totalDeletingCoordinatorDoc{0};
 };
 
+class TransactionCoordinatorsSSS final : public ServerStatusSection {
+public:
+    TransactionCoordinatorsSSS() : ServerStatusSection("twoPhaseCommitCoordinator") {}
+
+    bool includeByDefault() const override {
+        return true;
+    }
+
+    BSONObj generateSection(OperationContext* opCtx,
+                            const BSONElement& configElement) const override;
+};
+
 }  // namespace mongo
diff --git a/src/mongo/db/s/transaction_coordinator_test.cpp b/src/mongo/db/s/transaction_coordinator_test.cpp
index 31ac6201fb7..215e315fc97 100644
--- a/src/mongo/db/s/transaction_coordinator_test.cpp
+++ b/src/mongo/db/s/transaction_coordinator_test.cpp
@@ -35,6 +35,7 @@
 #include "mongo/db/commands/txn_cmds_gen.h"
 #include "mongo/db/commands/txn_two_phase_commit_cmds_gen.h"
 #include "mongo/db/dbdirectclient.h"
+#include "mongo/db/s/server_transaction_coordinators_metrics.h"
 #include "mongo/db/s/transaction_coordinator_document_gen.h"
 #include "mongo/db/s/transaction_coordinator_metrics_observer.h"
 #include "mongo/db/s/transaction_coordinator_test_fixture.h"
@@ -971,6 +972,28 @@ public:
                   metrics()->getCurrentDeletingCoordinatorDoc());
     }
 
+    void checkServerStatus() {
+        TransactionCoordinatorsSSS tcsss;
+        BSONElement dummy;
+        const auto serverStatusSection = tcsss.generateSection(operationContext(), dummy);
+        ASSERT_EQ(metrics()->getTotalCreated(), serverStatusSection["totalCreated"].Long());
+        ASSERT_EQ(metrics()->getTotalStartedTwoPhaseCommit(),
+                  serverStatusSection["totalStartedTwoPhaseCommit"].Long());
+        ASSERT_EQ(
+            metrics()->getCurrentWritingParticipantList(),
+            serverStatusSection.getObjectField("currentInSteps")["writingParticipantList"].Long());
+        ASSERT_EQ(metrics()->getCurrentWaitingForVotes(),
+                  serverStatusSection.getObjectField("currentInSteps")["waitingForVotes"].Long());
+        ASSERT_EQ(metrics()->getCurrentWritingDecision(),
+                  serverStatusSection.getObjectField("currentInSteps")["writingDecision"].Long());
+        ASSERT_EQ(
+            metrics()->getCurrentWaitingForDecisionAcks(),
+            serverStatusSection.getObjectField("currentInSteps")["waitingForDecisionAcks"].Long());
+        ASSERT_EQ(
+            metrics()->getCurrentDeletingCoordinatorDoc(),
+            serverStatusSection.getObjectField("currentInSteps")["deletingCoordinatorDoc"].Long());
+    }
+
     Date_t advanceClockSourceAndReturnNewNow() {
         const auto newNow = Date_t::now();
         clockSource()->reset(newNow);
@@ -2101,5 +2124,43 @@ TEST_F(TransactionCoordinatorMetricsTest, SlowLogLineIncludesStepDurationsAndTot
     ASSERT_EQUALS(1, countLogLinesContaining(" 500ms\n") + countLogLinesContaining(" 500ms\r\n"));
 }
 
+TEST_F(TransactionCoordinatorMetricsTest, ServerStatusSectionIncludesTotalCreated) {
+    metrics()->incrementTotalCreated();
+    checkServerStatus();
+}
+
+TEST_F(TransactionCoordinatorMetricsTest, ServerStatusSectionIncludesTotalStartedTwoPhaseCommit) {
+    metrics()->incrementTotalStartedTwoPhaseCommit();
+    checkServerStatus();
+}
+
+TEST_F(TransactionCoordinatorMetricsTest,
+       ServerStatusSectionIncludesCurrentWritingParticipantList) {
+    metrics()->incrementCurrentWritingParticipantList();
+    checkServerStatus();
+}
+
+TEST_F(TransactionCoordinatorMetricsTest, ServerStatusSectionIncludesCurrentWaitingForVotes) {
+    metrics()->incrementCurrentWaitingForVotes();
+    checkServerStatus();
+}
+
+TEST_F(TransactionCoordinatorMetricsTest, ServerStatusSectionIncludesCurrentWritingDecision) {
+    metrics()->incrementCurrentWritingDecision();
+    checkServerStatus();
+}
+
+TEST_F(TransactionCoordinatorMetricsTest,
+       ServerStatusSectionIncludesCurrentWaitingForDecisionAcks) {
+    metrics()->incrementCurrentWaitingForDecisionAcks();
+    checkServerStatus();
+}
+
+TEST_F(TransactionCoordinatorMetricsTest,
+       ServerStatusSectionIncludesCurrentDeletingCoordinatorDoc) {
+    metrics()->incrementCurrentDeletingCoordinatorDoc();
+    checkServerStatus();
+}
+
 }  // namespace
 }  // namespace mongo
diff --git a/src/mongo/db/s/transaction_coordinators_stats.idl b/src/mongo/db/s/transaction_coordinators_stats.idl
new file mode 100644
index 00000000000..7b1a9132536
--- /dev/null
+++ b/src/mongo/db/s/transaction_coordinators_stats.idl
@@ -0,0 +1,70 @@
+# Copyright (C) 2019-present MongoDB, Inc.
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the Server Side Public License, version 1,
+# as published by MongoDB, Inc.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# Server Side Public License for more details.
+#
+# You should have received a copy of the Server Side Public License
+# along with this program. If not, see
+# <http://www.mongodb.com/licensing/server-side-public-license>.
+#
+# As a special exception, the copyright holders give permission to link the
+# code of portions of this program with the OpenSSL library under certain
+# conditions as described in each individual source file and distribute
+# linked combinations including the program with the OpenSSL library. You
+# must comply with the Server Side Public License in all respects for
+# all of the code used other than as permitted herein. If you modify file(s)
+# with this exception, you may extend this exception to your version of the
+# file(s), but you are not obligated to do so. If you do not wish to do so,
+# delete this exception statement from your version. If you delete this
+# exception statement from all source files in the program, then also delete
+# it in the license file.
+#
+
+global:
+  cpp_namespace: "mongo"
+
+imports:
+  - "mongo/idl/basic_types.idl"
+
+structs:
+
+  CurrentInSteps:
+    description: "A struct describing the number of transaction coordinators currently in each
+                  two-phase commit step"
+    strict: true
+    fields:
+      writingParticipantList:
+        type: long
+        default: 0
+      waitingForVotes:
+        type: long
+        default: 0
+      writingDecision:
+        type: long
+        default: 0
+      waitingForDecisionAcks:
+        type: long
+        default: 0
+      deletingCoordinatorDoc:
+        type: long
+        default: 0
+
+  TransactionCoordinatorsStats:
+    description: "A struct representing the section of the server status
+                  command with information about transaction coordinators"
+    strict: true
+    fields:
+      totalCreated:
+        type: long
+        default: 0
+      totalStartedTwoPhaseCommit:
+        type: long
+        default: 0
+      currentInSteps:
+        type: CurrentInSteps