fix: use atomic.Int64 for workspace traffic metrics (#24844)

connMetrics.total was written to via atomic.AddInt64 and read as a plain
int64, producing a data race. Fix by switching the field to atomic.Int64
and using its typed Add/Load methods.

The testMetrics mock had a similar issue where mutex use was missing
where GetTotalBytes read the total bytes, which is also fixed in this
change.

scaletest/workspacetraffic/metrics.go

@@ -86,7 +86,7 @@ type connMetrics struct {
 	addError       func(float64)
 	observeLatency func(float64)
 	addTotal       func(float64)
-	total          int64
+	total          atomic.Int64
 }
 
 func (c *connMetrics) AddError(f float64) {
@@ -98,10 +98,10 @@ func (c *connMetrics) ObserveLatency(f float64) {
 }
 
 func (c *connMetrics) AddTotal(f float64) {
-	atomic.AddInt64(&c.total, int64(f))
+	c.total.Add(int64(f))
 	c.addTotal(f)
 }
 
 func (c *connMetrics) GetTotalBytes() int64 {
-	return c.total
+	return c.total.Load()
 }

scaletest/workspacetraffic/metrics_test.go

@@ -0,0 +1,48 @@
+package workspacetraffic_test
+
+import (
+	"sync"
+	"testing"
+
+	"github.com/prometheus/client_golang/prometheus"
+	"github.com/stretchr/testify/require"
+
+	"github.com/coder/coder/v2/scaletest/workspacetraffic"
+)
+
+func TestConnMetrics_Concurrent(t *testing.T) {
+	t.Parallel()
+
+	reg := prometheus.NewRegistry()
+	m := workspacetraffic.NewMetrics(reg, "username", "workspace_name", "agent_name")
+	cm := m.ReadMetrics("username", "workspace_name", "agent_name")
+
+	const (
+		writers       = 8
+		readers       = 8
+		opsPerWriter  = 1000
+		bytesPerWrite = 1
+	)
+
+	var wg sync.WaitGroup
+	wg.Add(writers + readers)
+	for i := 0; i < writers; i++ {
+		go func() {
+			defer wg.Done()
+			for j := 0; j < opsPerWriter; j++ {
+				cm.AddTotal(float64(bytesPerWrite))
+			}
+		}()
+	}
+	for i := 0; i < readers; i++ {
+		go func() {
+			defer wg.Done()
+			for j := 0; j < opsPerWriter; j++ {
+				_ = cm.GetTotalBytes()
+			}
+		}()
+	}
+	wg.Wait()
+
+	require.Equal(t, int64(writers*opsPerWriter*bytesPerWrite), cm.GetTotalBytes())
+}

scaletest/workspacetraffic/run_test.go

@@ -423,5 +423,7 @@ func (m *testMetrics) Latencies() []float64 {
 }
 
 func (m *testMetrics) GetTotalBytes() int64 {
+	m.Lock()
+	defer m.Unlock()
 	return int64(m.total)
 }