sql: report contention on writes in EXPLAIN ANALYZE

pkg/sql/opt/exec/explain/BUILD.bazel

@@ -60,6 +60,8 @@ go_test(
     embed = [":explain"],
     deps = [
         "//pkg/base",
+        "//pkg/security/securityassets",
+        "//pkg/security/securitytest",
         "//pkg/server",
         "//pkg/sql/catalog/colinfo",
         "//pkg/sql/execinfra",

pkg/sql/opt/exec/explain/main_test.go

@@ -14,11 +14,14 @@ import (
 	"os"
 	"testing"
 
+	"github.com/cockroachdb/cockroach/pkg/security/securityassets"
+	"github.com/cockroachdb/cockroach/pkg/security/securitytest"
 	"github.com/cockroachdb/cockroach/pkg/server"
 	"github.com/cockroachdb/cockroach/pkg/testutils/serverutils"
 )
 
 func TestMain(m *testing.M) {
+	securityassets.SetLoader(securitytest.EmbeddedAssets)
 	serverutils.InitTestServerFactory(server.TestServerFactory)
 	os.Exit(m.Run())
 }

pkg/sql/opt/exec/explain/output_test.go

@@ -19,14 +19,12 @@ import (
 	"time"
 
 	"github.com/cockroachdb/cockroach/pkg/base"
-	"github.com/cockroachdb/cockroach/pkg/server"
 	"github.com/cockroachdb/cockroach/pkg/sql/catalog/colinfo"
 	"github.com/cockroachdb/cockroach/pkg/sql/execinfra"
 	"github.com/cockroachdb/cockroach/pkg/sql/opt/exec/explain"
 	"github.com/cockroachdb/cockroach/pkg/sql/types"
 	"github.com/cockroachdb/cockroach/pkg/testutils"
 	"github.com/cockroachdb/cockroach/pkg/testutils/datapathutils"
-	"github.com/cockroachdb/cockroach/pkg/testutils/serverutils"
 	"github.com/cockroachdb/cockroach/pkg/testutils/skip"
 	"github.com/cockroachdb/cockroach/pkg/testutils/sqlutils"
 	"github.com/cockroachdb/cockroach/pkg/testutils/testcluster"
@@ -130,8 +128,6 @@ func TestMaxDiskSpillUsage(t *testing.T) {
 	distSQLKnobs := &execinfra.TestingKnobs{}
 	distSQLKnobs.ForceDiskSpill = true
 	testClusterArgs.ServerArgs.Knobs.DistSQL = distSQLKnobs
-	testClusterArgs.ServerArgs.Insecure = true
-	serverutils.InitTestServerFactory(server.TestServerFactory)
 	tc := testcluster.StartTestCluster(t, 1, testClusterArgs)
 	ctx := context.Background()
 	defer tc.Stopper().Stop(ctx)
@@ -150,9 +146,6 @@ CREATE TABLE t (a PRIMARY KEY, b) AS SELECT i, i FROM generate_series(1, 10) AS
 		for rows.Next() {
 			var res string
 			assert.NoError(t, rows.Scan(&res))
-			var sb strings.Builder
-			sb.WriteString(res)
-			sb.WriteByte('\n')
 			if matches := re.FindStringSubmatch(res); len(matches) > 0 {
 				return true
 			}
@@ -184,10 +177,8 @@ func TestCPUTimeEndToEnd(t *testing.T) {
 	distSQLKnobs := &execinfra.TestingKnobs{}
 	distSQLKnobs.ForceDiskSpill = true
 	testClusterArgs.ServerArgs.Knobs.DistSQL = distSQLKnobs
-	testClusterArgs.ServerArgs.Insecure = true
 	const numNodes = 3
 
-	serverutils.InitTestServerFactory(server.TestServerFactory)
 	tc := testcluster.StartTestCluster(t, numNodes, testClusterArgs)
 	ctx := context.Background()
 	defer tc.Stopper().Stop(ctx)
@@ -245,3 +236,123 @@ func TestCPUTimeEndToEnd(t *testing.T) {
 	runQuery("SELECT * FROM (VALUES (1), (2), (3)) v(a) INNER LOOKUP JOIN t ON a = x", false /* hideCPU */)
 	runQuery("SELECT count(*) FROM generate_series(1, 100000)", false /* hideCPU */)
 }
+
+// TestContentionTimeOnWrites verifies that the contention encountered during a
+// mutation is reported on EXPLAIN ANALYZE output.
+func TestContentionTimeOnWrites(t *testing.T) {
+	defer leaktest.AfterTest(t)()
+	defer log.Scope(t).Close(t)
+
+	tc := testcluster.StartTestCluster(t, 1, base.TestClusterArgs{})
+	ctx := context.Background()
+	defer tc.Stopper().Stop(ctx)
+
+	runner := sqlutils.MakeSQLRunner(tc.Conns[0])
+	runner.Exec(t, "CREATE TABLE t (k INT PRIMARY KEY, v INT)")
+
+	// The test involves three goroutines:
+	// - the main goroutine acts as the coordinator. It first waits for worker 1
+	//   to perform its mutation in an open txn, then blocks until worker 2
+	//   begins executing its mutation, then unblocks worker 1 and waits for
+	//   both workers to exit.
+	// - worker 1 goroutine performs a mutation without committing a txn. It
+	//   notifies the main goroutine by closing 'sem' once the mutation has been
+	//   performed. It then blocks until 'commitCh' is closed by the main
+	//   goroutine which allows worker 2 to experience contention.
+	// - worker 2 goroutine performs a mutation via EXPLAIN ANALYZE. This query
+	//   will be blocked until worker 1 commits its txn, so it should see
+	//   contention time reported in the output.
+
+	sem := make(chan struct{})
+	errCh := make(chan error, 1)
+	commitCh := make(chan struct{})
+	go func() {
+		defer close(errCh)
+		// Ensure that sem is always closed (in case we encounter an error
+		// before the mutation is performed).
+		var closedSem bool
+		defer func() {
+			if !closedSem {
+				close(sem)
+			}
+		}()
+		txn, err := tc.Conns[0].Begin()
+		if err != nil {
+			errCh <- err
+			return
+		}
+		_, err = txn.Exec("INSERT INTO t VALUES (1, 1)")
+		if err != nil {
+			errCh <- err
+			return
+		}
+		// Notify the main goroutine that the mutation has been performed.
+		close(sem)
+		closedSem = true
+		// Block until the main goroutine tells us that we're good to commit.
+		<-commitCh
+		if err = txn.Commit(); err != nil {
+			errCh <- err
+			return
+		}
+	}()
+
+	// Block until the mutation of worker 1 is done.
+	<-sem
+	// Check that no error was encountered before that.
+	select {
+	case err := <-errCh:
+		t.Fatal(err)
+	default:
+	}
+
+	var foundContention bool
+	errCh2 := make(chan error, 1)
+	go func() {
+		defer close(errCh2)
+		// Execute the mutation via EXPLAIN ANALYZE and check whether the
+		// contention is reported.
+		contentionRE := regexp.MustCompile(`cumulative time spent due to contention.*`)
+		rows := runner.Query(t, "EXPLAIN ANALYZE UPSERT INTO t VALUES (1, 2)")
+		for rows.Next() {
+			var line string
+			if err := rows.Scan(&line); err != nil {
+				errCh2 <- err
+				return
+			}
+			if contentionRE.MatchString(line) {
+				foundContention = true
+			}
+		}
+	}()
+
+	// Continuously poll the cluster queries until we see that the query that
+	// should be experiencing contention has started executing.
+	for {
+		row := runner.QueryRow(t, "SELECT count(*) FROM [SHOW CLUSTER QUERIES] WHERE query LIKE '%EXPLAIN ANALYZE UPSERT%'")
+		var count int
+		row.Scan(&count)
+		// Sleep for non-trivial amount of time to allow for worker 2 to start
+		// (if it hasn't already) and to experience the contention (if it has
+		// started).
+		time.Sleep(time.Second)
+		if count == 2 {
+			// We stop polling once we see 2 queries matching the LIKE pattern:
+			// the mutation query from worker 2 and the polling query itself.
+			break
+		}
+	}
+
+	// Allow worker 1 to commit which should unblock both workers.
+	close(commitCh)
+
+	// Wait for both workers to exit. Also perform sanity checks that the
+	// workers didn't run into any errors.
+	err := <-errCh
+	require.NoError(t, err)
+	err = <-errCh2
+	require.NoError(t, err)
+
+	// Meat of the test - verify that the contention was reported.
+	require.True(t, foundContention)
+}

pkg/sql/plan_node_to_row_source.go

@@ -47,6 +47,7 @@ type planNodeToRowSource struct {
 	// run time state machine values
 	row rowenc.EncDatumRow
 
+	contentionEventsListener  execstats.ContentionEventsListener
 	tenantConsumptionListener execstats.TenantConsumptionListener
 }
 
@@ -168,7 +169,7 @@ func (p *planNodeToRowSource) SetInput(ctx context.Context, input execinfra.RowS
 }
 
 func (p *planNodeToRowSource) Start(ctx context.Context) {
-	ctx = p.StartInternal(ctx, nodeName(p.node), &p.tenantConsumptionListener)
+	ctx = p.StartInternal(ctx, nodeName(p.node), &p.contentionEventsListener, &p.tenantConsumptionListener)
 	p.params.ctx = ctx
 	// This starts all of the nodes below this node.
 	if err := startExec(p.params, p.node); err != nil {
@@ -260,13 +261,14 @@ func (p *planNodeToRowSource) trailingMetaCallback() []execinfrapb.ProducerMetad
 
 // execStatsForTrace implements ProcessorBase.ExecStatsForTrace.
 func (p *planNodeToRowSource) execStatsForTrace() *execinfrapb.ComponentStats {
-	// Propagate RUs from IO requests.
-	// TODO(drewk): we should consider propagating other stats for planNode
-	// operators.
-	if p.tenantConsumptionListener.ConsumedRU == 0 {
+	// Propagate contention time and RUs from IO requests.
+	if p.contentionEventsListener.CumulativeContentionTime == 0 && p.tenantConsumptionListener.ConsumedRU == 0 {
 		return nil
 	}
 	return &execinfrapb.ComponentStats{
+		KV: execinfrapb.KVStats{
+			ContentionTime: optional.MakeTimeValue(p.contentionEventsListener.CumulativeContentionTime),
+		},
 		Exec: execinfrapb.ExecStats{
 			ConsumedRU: optional.MakeUint(p.tenantConsumptionListener.ConsumedRU),
 		},