Remove duplicated alert testing stack

This is confusing but this commit removes two alert tests:

[sig-instrumentation][Late] Alerts shouldn't report any unexpected alerts in firing or pending state [apigroup:config.openshift.io] [Suite:openshift/conformance/parallel]
- Ginko test run Late during e2e, live queries prometheus for the duration of
  the run, skips any alerts that have their own test, and flakes if any
  are found. It cannot fail.
- Flake rate is about 80%.
- This test is being removed in favor of the new one based on intervals
  observed. (see
  pkg/monitortests/testframework/legacytestframeworkmonitortests/monitortest.go.

[sig-arch] Check if alerts are firing during or after upgrade success
- Monitor test run after upgrade.
- Uses basically the same approach/code as above.
- Flake rate is 97%.
- This test is being removed in favor of the new one based on intervals
  observed. (see
  pkg/monitortests/testframework/legacytestframeworkmonitortests/monitortest.go.

Removing these two tests dramatically simplies our stack so we can begin
working on resurrecting some form of alert testing in just one spot.

The remaining backstop test is: [sig-trt][invariant] No alerts without an explicit test should be firing/pending more than historically
It is a 100% flake.

NOTE: A similarly named test that can fail and does catch real problems
is NOT being removed:

[sig-instrumentation] Prometheus [apigroup:image.openshift.io] when installed on the cluster shouldn't report any alerts in firing state apart from Watchdog and AlertmanagerReceiversNotConfigured [Early][apigroup:config.openshift.io] [Skipped:Disconnected] [Suite:openshift/conformance/parallel]

Author: dgoodwin

pkg/alerts/check.go

@@ -12,7 +12,6 @@ import (
 	configv1 "github.com/openshift/api/config/v1"
 	"github.com/sirupsen/logrus"
 
-	"github.com/openshift/origin/pkg/alerts"
 	"github.com/openshift/origin/pkg/monitor/monitorapi"
 	"github.com/openshift/origin/pkg/test/ginkgo/junitapi"
 	helper "github.com/openshift/origin/test/extended/util/prometheus"
@@ -26,8 +25,10 @@ import (
 	"k8s.io/kubernetes/test/e2e/framework"
 )
 
+type AllowedAlertsFunc func(featureSet configv1.FeatureSet) (allowedFiringWithBugs, allowedFiring, allowedPendingWithBugs, allowedPending helper.MetricConditions)
+
 func testAlerts(events monitorapi.Intervals,
-	allowancesFunc alerts.AllowedAlertsFunc,
+	allowancesFunc AllowedAlertsFunc,
 	jobType *platformidentification.JobType,
 	restConfig *rest.Config,
 	duration time.Duration,
@@ -74,7 +75,7 @@ func testAlerts(events monitorapi.Intervals,
 }
 
 func RunAlertTests(jobType *platformidentification.JobType,
-	allowancesFunc alerts.AllowedAlertsFunc,
+	allowancesFunc AllowedAlertsFunc,
 	featureSet configv1.FeatureSet,
 	etcdAllowance allowedalerts2.AlertTestAllowanceCalculator,
 	events monitorapi.Intervals,
@@ -109,7 +110,7 @@ func RunAlertTests(jobType *platformidentification.JobType,
 // runBackstopTest will process the intervals for any alerts which do not have their own explicit test,
 // and look for any pending/firing intervals that are not within sufficient range.
 func runBackstopTest(
-	allowancesFunc alerts.AllowedAlertsFunc,
+	allowancesFunc AllowedAlertsFunc,
 	featureSet configv1.FeatureSet,
 	alertIntervals monitorapi.Intervals,
 	alertTests []allowedalerts2.AlertTest) []*junitapi.JUnitTestCase {

pkg/monitortests/testframework/legacytestframeworkmonitortests/alerts.go

@@ -38,7 +38,6 @@ import (
 	"k8s.io/kubernetes/test/e2e/upgrades/node"
 
 	"github.com/openshift/origin/test/e2e/upgrade/adminack"
-	"github.com/openshift/origin/test/e2e/upgrade/alert"
 	"github.com/openshift/origin/test/e2e/upgrade/dns"
 	"github.com/openshift/origin/test/e2e/upgrade/manifestdelete"
 	"github.com/openshift/origin/test/extended/prometheus"
@@ -56,7 +55,6 @@ func AllTests() []upgrades.Test {
 	return []upgrades.Test{
 		&adminack.UpgradeTest{},
 		&manifestdelete.UpgradeTest{},
-		&alert.UpgradeTest{},
 
 		&node.SecretUpgradeTest{},
 		&apps.ReplicaSetUpgradeTest{},

test/e2e/upgrade/alert/alert.go

@@ -16,7 +16,6 @@ import (
 
 	g "github.com/onsi/ginkgo/v2"
 	o "github.com/onsi/gomega"
-	"github.com/openshift/origin/pkg/alerts"
 	promv1 "github.com/prometheus/client_golang/api/prometheus/v1"
 	dto "github.com/prometheus/client_model/go"
 	"github.com/prometheus/common/expfmt"
@@ -252,12 +251,6 @@ var _ = g.Describe("[sig-instrumentation][Late] Alerts", func() {
 		}
 	})
 
-	g.It("shouldn't report any unexpected alerts in firing or pending state", func() {
-		// we only consider samples since the beginning of the test
-		testDuration := exutil.DurationSinceStartInSeconds()
-		alerts.CheckAlerts(alerts.AllowedAlertsDuringConformance, oc.NewPrometheusClient(context.TODO()), oc.AdminConfigClient(), testDuration, nil)
-	})
-
 	g.It("shouldn't exceed the series limit of total series sent via telemetry from each cluster", func() {
 		if enabledErr, err := telemetryIsEnabled(ctx, oc.AdminKubeClient()); err != nil {
 			e2e.Failf("could not determine if Telemetry is enabled: %v", err)