From 316b19a5d164fd508e53b46d637f326bceb1cdf0 Mon Sep 17 00:00:00 2001
From: Roman Khavronenko <roman@victoriametrics.com>
Date: Fri, 19 Apr 2024 09:06:40 +0200
Subject: [PATCH] app/vmalert: make `TestGroupStart` more reliable (#6130)

There was a sleep statement in the test, waiting for Group
to perform a couple of evaluation. But looks like
it worked unreliable for some CI tests like the one below
https://github.com/VictoriaMetrics/VictoriaMetrics/actions/runs/8718213844/job/23915007958?pr=6115

This commit changes the sleep statement on a function that
waits for a specific number of evaluations. It should make this
test faster in general case, and more reliable for slow environemnts.
---
 app/vmalert/rule/group_test.go | 26 +++++++++++++++++++++-----
 1 file changed, 21 insertions(+), 5 deletions(-)

diff --git a/app/vmalert/rule/group_test.go b/app/vmalert/rule/group_test.go
index d1777b6c5..b7560c1ae 100644
--- a/app/vmalert/rule/group_test.go
+++ b/app/vmalert/rule/group_test.go
@@ -217,7 +217,6 @@ func TestGroupStart(t *testing.T) {
 
 	const evalInterval = time.Millisecond
 	g := NewGroup(groups[0], fs, evalInterval, map[string]string{"cluster": "east-1"})
-	g.Concurrency = 2
 
 	const inst1, inst2, job = "foo", "bar", "baz"
 	m1 := metricWithLabels(t, "instance", inst1, "job", job)
@@ -262,8 +261,25 @@ func TestGroupStart(t *testing.T) {
 		close(finished)
 	}()
 
-	// wait for multiple evals
-	time.Sleep(20 * evalInterval)
+	waitForIterations := func(n int, interval time.Duration) {
+		t.Helper()
+
+		var cur uint64
+		prev := g.metrics.iterationTotal.Get()
+		for i := 0; ; i++ {
+			if i > 40 {
+				t.Fatalf("group wasn't able to perform %d evaluations during %d eval intervals", n, i)
+			}
+			cur = g.metrics.iterationTotal.Get()
+			if int(cur-prev) >= n {
+				return
+			}
+			time.Sleep(interval)
+		}
+	}
+
+	// wait for multiple evaluation iterations
+	waitForIterations(4, evalInterval)
 
 	gotAlerts := fn.GetAlerts()
 	expectedAlerts := []notifier.Alert{*alert1, *alert2}
@@ -280,8 +296,8 @@ func TestGroupStart(t *testing.T) {
 	// and set only one datapoint for response
 	fs.Add(m1)
 
-	// wait for multiple evals
-	time.Sleep(20 * evalInterval)
+	// wait for multiple evaluation iterations
+	waitForIterations(4, evalInterval)
 
 	gotAlerts = fn.GetAlerts()
 	alert2.State = notifier.StateInactive