Fix linter issues and slightly improve some tests

output/cloud/data.go

@@ -245,7 +245,9 @@ func (d durations) Less(i, j int) bool { return d[i] < d[j] }
 // Used when there are fewer samples in the bucket (so we can interpolate)
 // and for benchmark comparisons and verification of the quickselect
 // algorithm (it should return exactly the same results if interpolation isn't used).
-func (d durations) SortGetNormalBounds(radius, iqrLowerCoef, iqrUpperCoef float64, interpolate bool) (min, max time.Duration) {
+func (d durations) SortGetNormalBounds(
+	radius, iqrLowerCoef, iqrUpperCoef float64, interpolate bool,
+) (min, max time.Duration) {
 	if len(d) == 0 {
 		return
 	}
@@ -276,7 +278,7 @@ func (d durations) SortGetNormalBounds(radius, iqrLowerCoef, iqrUpperCoef float6
 
 	min = q1 - time.Duration(iqrLowerCoef*iqr) // lower fence, anything below this is an outlier
 	max = q3 + time.Duration(iqrUpperCoef*iqr) // upper fence, anything above this is an outlier
-	return
+	return min, max
 }
 
 // Reworked and translated in Go from:
@@ -288,7 +290,7 @@ func (d durations) SortGetNormalBounds(radius, iqrLowerCoef, iqrUpperCoef float6
 // that only depends on the sort.Interface methods, but that would
 // probably introduce some performance overhead because of the
 // dynamic dispatch.
-func (d durations) quickSelect(k int) time.Duration {
+func (d durations) quickSelect(k int) time.Duration { //nolint:gocognit
 	n := len(d)
 	l := 0
 	ir := n - 1

output/cloud/data_test.go

@@ -233,10 +233,10 @@ func TestMetricAggregation(t *testing.T) {
 // I've not used that after the initial tests because it's a big
 // external dependency that's not really needed for the tests at
 // this point.
-func getDurations(count int, min, multiplier float64) durations {
+func getDurations(r *rand.Rand, count int, min, multiplier float64) durations {
 	data := make(durations, count)
 	for j := 0; j < count; j++ {
-		data[j] = time.Duration(min + rand.Float64()*multiplier)
+		data[j] = time.Duration(min + r.Float64()*multiplier) //nolint:gosec
 	}
 	return data
 }
@@ -246,10 +246,14 @@ func BenchmarkDurationBounds(b *testing.B) {
 	iqrLowerCoef := 1.5
 	iqrUpperCoef := 1.5
 
+	seed := time.Now().UnixNano()
+	r := rand.New(rand.NewSource(seed)) //nolint:gosec
+	b.Logf("Random source seeded with %d\n", seed)
+
 	getData := func(b *testing.B, count int) durations {
 		b.StopTimer()
 		defer b.StartTimer()
-		return getDurations(count, 0.1*float64(time.Second), float64(time.Second))
+		return getDurations(r, count, 0.1*float64(time.Second), float64(time.Second))
 	}
 
 	for count := 100; count <= 5000; count += 500 {
@@ -277,11 +281,15 @@ func BenchmarkDurationBounds(b *testing.B) {
 
 func TestQuickSelectAndBounds(t *testing.T) {
 	t.Parallel()
+
+	seed := time.Now().UnixNano()
+	r := rand.New(rand.NewSource(seed)) //nolint:gosec
+	t.Logf("Random source seeded with %d\n", seed)
+
 	mult := time.Millisecond
-	for _, count := range []int{1, 2, 3, 4, 5, 10, 15, 20, 25, 50, 100, 250 + rand.Intn(100)} {
+	for _, count := range []int{1, 2, 3, 4, 5, 10, 15, 20, 25, 50, 100, 250 + r.Intn(100)} {
 		count := count
 		t.Run(fmt.Sprintf("simple-%d", count), func(t *testing.T) {
-			t.Parallel()
 			data := make(durations, count)
 			for i := 0; i < count; i++ {
 				data[i] = time.Duration(i) * mult
@@ -290,13 +298,11 @@ func TestQuickSelectAndBounds(t *testing.T) {
 			for i := 0; i < 10; i++ {
 				dataCopy := make(durations, count)
 				assert.Equal(t, count, copy(dataCopy, data))
-				k := rand.Intn(count)
+				k := r.Intn(count)
 				assert.Equal(t, dataCopy.quickSelect(k), time.Duration(k)*mult)
 			}
 		})
 		t.Run(fmt.Sprintf("random-%d", count), func(t *testing.T) {
-			t.Parallel()
-
 			testCases := []struct{ r, l, u float64 }{
 				{0.25, 1.5, 1.5}, // Textbook
 				{0.25, 1.5, 1.3}, // Defaults
@@ -306,7 +312,7 @@ func TestQuickSelectAndBounds(t *testing.T) {
 
 			for tcNum, tc := range testCases {
 				tc := tc
-				data := getDurations(count, 0.3*float64(time.Second), 2*float64(time.Second))
+				data := getDurations(r, count, 0.3*float64(time.Second), 2*float64(time.Second))
 				dataForSort := make(durations, count)
 				dataForSelect := make(durations, count)
 				assert.Equal(t, count, copy(dataForSort, data))
@@ -320,7 +326,7 @@ func TestQuickSelectAndBounds(t *testing.T) {
 					assert.Equal(t, sortMin, selectMin)
 					assert.Equal(t, sortMax, selectMax)
 
-					k := rand.Intn(count)
+					k := r.Intn(count)
 					assert.Equal(t, dataForSort[k], dataForSelect.quickSelect(k))
 					assert.Equal(t, dataForSort[k], data.quickSelect(k))
 				})

output/cloud/output_test.go

@@ -126,8 +126,8 @@ func getSampleChecker(t *testing.T, expSamples <-chan []Sample) http.HandlerFunc
 	}
 }
 
-func skewTrail(t httpext.Trail, minCoef, maxCoef float64) httpext.Trail {
-	coef := minCoef + rand.Float64()*(maxCoef-minCoef)
+func skewTrail(r *rand.Rand, t httpext.Trail, minCoef, maxCoef float64) httpext.Trail {
+	coef := minCoef + r.Float64()*(maxCoef-minCoef)
 	addJitter := func(d *time.Duration) {
 		*d = time.Duration(float64(*d) * coef)
 	}
@@ -158,6 +158,10 @@ func TestCloudOutput(t *testing.T) {
 }
 
 func runCloudOutputTestCase(t *testing.T, minSamples int) {
+	seed := time.Now().UnixNano()
+	r := rand.New(rand.NewSource(seed)) //nolint:gosec
+	t.Logf("Random source seeded with %d\n", seed)
+
 	tb := httpmultibin.NewHTTPMultiBin(t)
 	tb.Mux.HandleFunc("/v1/tests", http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
 		_, err := fmt.Fprintf(w, `{
@@ -254,7 +258,7 @@ func runCloudOutputTestCase(t *testing.T, minSamples int) {
 	trails := []stats.SampleContainer{}
 	durations := make([]time.Duration, len(trails))
 	for i := int64(0); i < out.config.AggregationMinSamples.Int64; i++ {
-		similarTrail := skewTrail(simpleTrail, 1.0, 1.0+smallSkew)
+		similarTrail := skewTrail(r, simpleTrail, 1.0, 1.0+smallSkew)
 		trails = append(trails, &similarTrail)
 		durations = append(durations, similarTrail.Duration)
 	}
@@ -269,7 +273,7 @@ func runCloudOutputTestCase(t *testing.T, minSamples int) {
 		assert.InEpsilon(t, normal, stats.ToD(aggr.Max), smallSkew)
 	}
 
-	outlierTrail := skewTrail(simpleTrail, 2.0+smallSkew, 3.0+smallSkew)
+	outlierTrail := skewTrail(r, simpleTrail, 2.0+smallSkew, 3.0+smallSkew)
 	trails = append(trails, &outlierTrail)
 	out.AddMetricSamples(trails)
 	expSamples <- []Sample{