fix: exit after a timeout on stuck iterations

nvloff-f3 · nvloff-f3 · commit b43bfb92e9c1 · 2024-07-12T09:43:39.000+03:00
In case an interation is stuck for whatever reason f1 will never exit.

This adds a 10s timeout to wait for active iterations to complete.
diff --git a/internal/run/run_cmd.go b/internal/run/run_cmd.go
@@ -16,6 +16,8 @@ import (
 	"github.com/form3tech-oss/f1/v2/pkg/f1/scenarios"
 )
 
+const waitForCompletionTimeout = 10 * time.Second
+
 func Cmd(
 	s *scenarios.Scenarios,
 	builders []api.Builder,
@@ -157,7 +159,7 @@ func runCmdExecute(
 			MaxFailures:     maxFailures,
 			MaxFailuresRate: maxFailuresRate,
 			IgnoreDropped:   ignoreDropped,
-		}, s, trig, settings, metricsInstance, output)
+		}, s, trig, waitForCompletionTimeout, settings, metricsInstance, output)
 		if err != nil {
 			return fmt.Errorf("new run: %w", err)
 		}
diff --git a/internal/run/run_cmd_test.go b/internal/run/run_cmd_test.go
@@ -563,7 +563,68 @@ func TestInterruptedRun(t *testing.T) {
 	then.
 		setup_teardown_is_called_within(600 * time.Millisecond).and().
 		metrics_are_pushed_to_prometheus().and().
-		there_is_a_metric_called("form3_loadtest_iteration")
+		there_is_a_metric_called("form3_loadtest_iteration").and().
+		expect_the_stdout_output_to_include([]string{
+			"Interrupted - waiting for active tests to complete",
+		})
+}
+
+func TestInterruptedRun_TimesOut(t *testing.T) {
+	t.Parallel()
+
+	given, when, then := NewRunTestStage(t)
+
+	given.
+		a_timer_is_started().
+		a_rate_of("1/1s").and().
+		a_duration_of(5 * time.Second).and().
+		a_scenario_where_each_iteration_takes(3 * time.Second).and().
+		wait_for_completion_timeout_of(1 * time.Second).and().
+		a_distribution_type("none")
+
+	when.
+		the_run_command_is_executed_and_cancelled_after(500 * time.Millisecond)
+
+	then.
+		setup_teardown_is_called_within(600*time.Millisecond + 1*time.Second).and().
+		metrics_are_pushed_to_prometheus().and().
+		expect_the_stdout_output_to_include([]string{
+			"Interrupted - waiting for active tests to complete",
+			"Active tests not completed after 1s. Stopping...",
+		})
+
+	// needed for goroutine checker
+	// sleep is the time left in the interation duration after the timeout
+	time.Sleep(3 * time.Second)
+}
+
+func TestMaxDurationReached_TimesOut(t *testing.T) {
+	t.Parallel()
+
+	given, when, then := NewRunTestStage(t)
+
+	given.
+		a_timer_is_started().
+		a_rate_of("1/1s").and().
+		a_duration_of(500 * time.Millisecond).and().
+		a_scenario_where_each_iteration_takes(3 * time.Second).and().
+		wait_for_completion_timeout_of(1 * time.Second).and().
+		a_distribution_type("none")
+
+	when.
+		the_run_command_is_executed()
+
+	then.
+		setup_teardown_is_called_within(600*time.Millisecond + 1*time.Second).and().
+		metrics_are_pushed_to_prometheus().and().
+		expect_the_stdout_output_to_include([]string{
+			"Max Duration Elapsed - waiting for active tests to complete",
+			"Active tests not completed after 1s. Stopping...",
+		})
+
+	// needed for goroutine checker
+	// sleep is the time left in the interation duration after the timeout
+	time.Sleep(3 * time.Second)
 }
 
 func TestFinalRunMetrics(t *testing.T) {
diff --git a/internal/run/run_stage_test.go b/internal/run/run_stage_test.go
@@ -68,58 +68,60 @@ type (
 )
 
 type RunTestStage struct {
-	startTime              time.Time
-	metrics                *metrics.Metrics
-	output                 *ui.Output
-	runInstance            *run.Run
-	runResult              *run.Result
-	t                      *testing.T
-	require                *require.Assertions
-	metricData             *MetricData
-	scenarioCleanup        func()
-	assert                 *assert.Assertions
-	iterationCleanup       func()
-	f1                     *f1.F1
-	durations              sync.Map
-	frequency              string
-	rate                   string
-	stages                 string
-	distributionType       string
-	configFile             string
-	startRate              string
-	endRate                string
-	rampDuration           string
-	scenario               string
-	settings               envsettings.Settings
-	maxFailures            uint64
-	maxIterations          uint64
-	maxFailuresRate        int
-	duration               time.Duration
-	concurrency            int
-	triggerType            TriggerType
-	iterationTeardownCount atomic.Uint32
-	setupTeardownCount     atomic.Uint32
-	runCount               atomic.Uint32
-	stdout                 syncWriter
-	stderr                 syncWriter
-	interactive            bool
-	verbose                bool
+	startTime                time.Time
+	metrics                  *metrics.Metrics
+	output                   *ui.Output
+	runInstance              *run.Run
+	runResult                *run.Result
+	t                        *testing.T
+	require                  *require.Assertions
+	metricData               *MetricData
+	scenarioCleanup          func()
+	assert                   *assert.Assertions
+	iterationCleanup         func()
+	f1                       *f1.F1
+	durations                sync.Map
+	frequency                string
+	rate                     string
+	stages                   string
+	distributionType         string
+	configFile               string
+	startRate                string
+	endRate                  string
+	rampDuration             string
+	scenario                 string
+	settings                 envsettings.Settings
+	maxFailures              uint64
+	maxIterations            uint64
+	maxFailuresRate          int
+	duration                 time.Duration
+	waitForCompletionTimeout time.Duration
+	concurrency              int
+	triggerType              TriggerType
+	iterationTeardownCount   atomic.Uint32
+	setupTeardownCount       atomic.Uint32
+	runCount                 atomic.Uint32
+	stdout                   syncWriter
+	stderr                   syncWriter
+	interactive              bool
+	verbose                  bool
 }
 
 func NewRunTestStage(t *testing.T) (*RunTestStage, *RunTestStage, *RunTestStage) {
 	t.Helper()
 	stage := &RunTestStage{
-		t:           t,
-		concurrency: 100,
-		assert:      assert.New(t),
-		require:     require.New(t),
-		f1:          f1.New(),
-		settings:    envsettings.Get(),
-		metricData:  NewMetricData(),
-		output:      ui.NewDiscardOutput(),
-		metrics:     metrics.NewInstance(prometheus.NewRegistry(), true),
-		stdout:      syncWriter{writer: &bytes.Buffer{}},
-		stderr:      syncWriter{writer: &bytes.Buffer{}},
+		t:                        t,
+		concurrency:              100,
+		assert:                   assert.New(t),
+		require:                  require.New(t),
+		f1:                       f1.New(),
+		settings:                 envsettings.Get(),
+		metricData:               NewMetricData(),
+		output:                   ui.NewDiscardOutput(),
+		metrics:                  metrics.NewInstance(prometheus.NewRegistry(), true),
+		stdout:                   syncWriter{writer: &bytes.Buffer{}},
+		stderr:                   syncWriter{writer: &bytes.Buffer{}},
+		waitForCompletionTimeout: 5 * time.Second,
 	}
 
 	handler := FakePrometheusHandler(t, stage.metricData)
@@ -141,6 +143,11 @@ func (s *RunTestStage) a_rate_of(rate string) *RunTestStage {
 	return s
 }
 
+func (s *RunTestStage) wait_for_completion_timeout_of(timeout time.Duration) *RunTestStage {
+	s.waitForCompletionTimeout = timeout
+	return s
+}
+
 func (s *RunTestStage) and() *RunTestStage {
 	return s
 }
@@ -198,7 +205,7 @@ func (s *RunTestStage) setupRun() {
 		MaxFailures:     s.maxFailures,
 		MaxFailuresRate: s.maxFailuresRate,
 		Verbose:         s.verbose,
-	}, s.f1.GetScenarios(), s.build_trigger(), s.settings, s.metrics, outputer)
+	}, s.f1.GetScenarios(), s.build_trigger(), s.waitForCompletionTimeout, s.settings, s.metrics, outputer)
 
 	s.require.NoError(err)
 	s.runInstance = r
@@ -218,8 +225,11 @@ func (s *RunTestStage) the_run_command_is_executed_and_cancelled_after(duration
 	s.setupRun()
 
 	var err error
-	ctx, cancel := context.WithTimeout(context.TODO(), duration)
-	defer cancel()
+	ctx, cancel := context.WithCancel(context.TODO())
+	go func() {
+		<-time.After(duration)
+		cancel()
+	}()
 
 	s.runResult, err = s.runInstance.Do(ctx)
 	s.require.NoError(err)
diff --git a/internal/run/test_runner.go b/internal/run/test_runner.go
@@ -30,22 +30,24 @@ const (
 )
 
 type Run struct {
-	pusher         *push.Pusher
-	progressRunner *raterun.Runner
-	metrics        *metrics.Metrics
-	views          *views.Views
-	activeScenario *workers.ActiveScenario
-	trigger        *api.Trigger
-	output         *ui.Output
-	scenarioLogger *ScenarioLogger
-	result         *Result
-	options        options.RunOptions
+	pusher                   *push.Pusher
+	progressRunner           *raterun.Runner
+	metrics                  *metrics.Metrics
+	views                    *views.Views
+	activeScenario           *workers.ActiveScenario
+	trigger                  *api.Trigger
+	output                   *ui.Output
+	scenarioLogger           *ScenarioLogger
+	result                   *Result
+	options                  options.RunOptions
+	waitForCompletionTimeout time.Duration
 }
 
 func NewRun(
 	options options.RunOptions,
 	scenarios *scenarios.Scenarios,
 	trigger *api.Trigger,
+	waitForCompletionTimeout time.Duration,
 	settings envsettings.Settings,
 	metricsInstance *metrics.Metrics,
 	parentOutput *ui.Output,
@@ -91,16 +93,17 @@ func NewRun(
 	pusher := newMetricsPusher(settings, scenario.Name, metricsInstance)
 
 	return &Run{
-		options:        options,
-		trigger:        trigger,
-		metrics:        metricsInstance,
-		views:          viewsInstance,
-		result:         result,
-		pusher:         pusher,
-		output:         outputer,
-		progressRunner: progressRunner,
-		activeScenario: activeScenario,
-		scenarioLogger: scenarioLogger,
+		options:                  options,
+		trigger:                  trigger,
+		metrics:                  metricsInstance,
+		views:                    viewsInstance,
+		result:                   result,
+		pusher:                   pusher,
+		output:                   outputer,
+		progressRunner:           progressRunner,
+		activeScenario:           activeScenario,
+		scenarioLogger:           scenarioLogger,
+		waitForCompletionTimeout: waitForCompletionTimeout,
 	}, nil
 }
 
@@ -254,12 +257,27 @@ func (r *Run) run(ctx context.Context) {
 	case <-ctx.Done():
 		r.output.Display(r.result.Interrupted())
 		r.progressRunner.Restart()
-		<-poolManager.WaitForCompletion()
+		select {
+		case <-poolManager.WaitForCompletion():
+		case <-time.After(r.waitForCompletionTimeout):
+			r.output.Display(ui.WarningMessage{
+				Message: fmt.Sprintf("Active tests not completed after %s. Stopping...", r.waitForCompletionTimeout.String()),
+			})
+		}
+
 	case <-triggerCtx.Done():
 		if triggerCtx.Err() == context.DeadlineExceeded {
 			r.output.Display(r.result.MaxDurationElapsed())
+		} else {
+			r.output.Display(r.result.Interrupted())
+		}
+		select {
+		case <-poolManager.WaitForCompletion():
+		case <-time.After(r.waitForCompletionTimeout):
+			r.output.Display(ui.WarningMessage{
+				Message: fmt.Sprintf("Active tests not completed after %s. Stopping...", r.waitForCompletionTimeout.String()),
+			})
 		}
-		<-poolManager.WaitForCompletion()
 	case <-poolManager.WaitForCompletion():
 		if poolManager.MaxIterationsReached() {
 			r.output.Display(r.result.MaxIterationsReached())
diff --git a/internal/ui/printer.go b/internal/ui/printer.go
@@ -34,14 +34,6 @@ func (t *Printer) Error(a ...any) {
 	fmt.Fprintln(t.ErrWriter, a...)
 }
 
-func (t *Printer) Printf(format string, a ...any) {
-	fmt.Fprintf(t.Writer, format, a...)
-}
-
 func (t *Printer) Warn(a ...any) {
-	fmt.Fprint(t.ErrWriter, a...)
-}
-
-func (t *Printer) Warnf(format string, a ...any) {
-	fmt.Fprintf(t.ErrWriter, format, a...)
+	fmt.Fprintln(t.ErrWriter, a...)
 }
