e2e: fix ZipkinTracing flaky

charts/gateway-addons-helm/README.md

@@ -61,6 +61,7 @@ To uninstall the chart:
 | fluent-bit.config.service | string | `"[SERVICE]\n    Daemon Off\n    Flush {{ .Values.flush }}\n    Log_Level {{ .Values.logLevel }}\n    Parsers_File parsers.conf\n    Parsers_File custom_parsers.conf\n    HTTP_Server On\n    HTTP_Listen 0.0.0.0\n    HTTP_Port {{ .Values.metricsPort }}\n    Health_Check On\n"` |  |
 | fluent-bit.enabled | bool | `true` |  |
 | fluent-bit.fullnameOverride | string | `"fluent-bit"` |  |
+| fluent-bit.image.repository | string | `"fluent/fluent-bit"` |  |
 | fluent-bit.podAnnotations."fluentbit.io/exclude" | string | `"true"` |  |
 | fluent-bit.podAnnotations."prometheus.io/path" | string | `"/api/v1/metrics/prometheus"` |  |
 | fluent-bit.podAnnotations."prometheus.io/port" | string | `"2020"` |  |

charts/gateway-addons-helm/values.yaml

@@ -60,6 +60,8 @@ prometheus:
 # Values for Fluent-bit dependency
 fluent-bit:
   enabled: true
+  image:
+    repository: fluent/fluent-bit # use image from dockerhub
   fullnameOverride: fluent-bit
   testFramework:
     enabled: false

site/content/en/latest/install/gateway-addons-helm-api.md

@@ -40,6 +40,7 @@ An Add-ons Helm chart for Envoy Gateway
 | fluent-bit.config.service | string | `"[SERVICE]\n    Daemon Off\n    Flush {{ .Values.flush }}\n    Log_Level {{ .Values.logLevel }}\n    Parsers_File parsers.conf\n    Parsers_File custom_parsers.conf\n    HTTP_Server On\n    HTTP_Listen 0.0.0.0\n    HTTP_Port {{ .Values.metricsPort }}\n    Health_Check On\n"` |  |
 | fluent-bit.enabled | bool | `true` |  |
 | fluent-bit.fullnameOverride | string | `"fluent-bit"` |  |
+| fluent-bit.image.repository | string | `"fluent/fluent-bit"` |  |
 | fluent-bit.podAnnotations."fluentbit.io/exclude" | string | `"true"` |  |
 | fluent-bit.podAnnotations."prometheus.io/path" | string | `"/api/v1/metrics/prometheus"` |  |
 | fluent-bit.podAnnotations."prometheus.io/port" | string | `"2020"` |  |

site/content/zh/latest/install/gateway-addons-helm-api.md

@@ -40,6 +40,7 @@ An Add-ons Helm chart for Envoy Gateway
 | fluent-bit.config.service | string | `"[SERVICE]\n    Daemon Off\n    Flush {{ .Values.flush }}\n    Log_Level {{ .Values.logLevel }}\n    Parsers_File parsers.conf\n    Parsers_File custom_parsers.conf\n    HTTP_Server On\n    HTTP_Listen 0.0.0.0\n    HTTP_Port {{ .Values.metricsPort }}\n    Health_Check On\n"` |  |
 | fluent-bit.enabled | bool | `true` |  |
 | fluent-bit.fullnameOverride | string | `"fluent-bit"` |  |
+| fluent-bit.image.repository | string | `"fluent/fluent-bit"` |  |
 | fluent-bit.podAnnotations."fluentbit.io/exclude" | string | `"true"` |  |
 | fluent-bit.podAnnotations."prometheus.io/path" | string | `"/api/v1/metrics/prometheus"` |  |
 | fluent-bit.podAnnotations."prometheus.io/port" | string | `"2020"` |  |

test/e2e/base/manifests.yaml

@@ -42,11 +42,6 @@ spec:
     allowedRoutes:
       namespaces:
         from: All
-  infrastructure:
-    parametersRef:
-      group: gateway.envoyproxy.io
-      kind: EnvoyProxy
-      name: zipkin-tracing
 ---
 apiVersion: v1
 kind: Service

test/e2e/tests/tracing.go

@@ -11,20 +11,11 @@ package tests
 import (
 	"context"
 	"fmt"
-	"net"
-	"net/http"
-	"net/url"
-	"strings"
 	"testing"
 	"time"
 
-	"github.com/go-logfmt/logfmt"
-	"github.com/gogo/protobuf/jsonpb" // nolint: depguard // tempopb use gogo/protobuf
-	"github.com/grafana/tempo/pkg/tempopb"
-	corev1 "k8s.io/api/core/v1"
 	"k8s.io/apimachinery/pkg/types"
 	"k8s.io/apimachinery/pkg/util/wait"
-	"sigs.k8s.io/controller-runtime/pkg/client"
 	httputils "sigs.k8s.io/gateway-api/conformance/utils/http"
 	"sigs.k8s.io/gateway-api/conformance/utils/kubernetes"
 	"sigs.k8s.io/gateway-api/conformance/utils/suite"
@@ -114,94 +105,39 @@ var ZipkinTracingTest = suite.ConformanceTest{
 				// should make them kept consistent
 				"service.name": fmt.Sprintf("%s/%s", gwNN.Namespace, gwNN.Name),
 			}
-			// let's wait for the log to be sent to stdout
 			if err := wait.PollUntilContextTimeout(context.TODO(), time.Second, time.Minute, true,
 				func(ctx context.Context) (bool, error) {
-					count, err := QueryTraceFromTempo(t, suite.Client, tags)
+					preCount, err := QueryTraceFromTempo(t, suite.Client, tags)
 					if err != nil {
 						t.Logf("failed to get trace count from tempo: %v", err)
 						return false, nil
 					}
 
-					if count > 0 {
-						return true, nil
+					httputils.MakeRequestAndExpectEventuallyConsistentResponse(t, suite.RoundTripper, suite.TimeoutConfig, gwAddr, expectedResponse)
+
+					// looks like we need almost 15 seconds to get the trace from Tempo?
+					err = wait.PollUntilContextTimeout(context.TODO(), time.Second, 15*time.Second, true, func(ctx context.Context) (done bool, err error) {
+						curCount, err := QueryTraceFromTempo(t, suite.Client, tags)
+						if err != nil {
+							t.Logf("failed to get curCount count from tempo: %v", err)
+							return false, nil
+						}
+
+						if curCount > preCount {
+							return true, nil
+						}
+
+						return false, nil
+					})
+					if err != nil {
+						t.Logf("failed to get current count from tempo: %v", err)
+						return false, nil
 					}
-					return false, nil
+
+					return true, nil
 				}); err != nil {
 				t.Errorf("failed to get trace from tempo: %v", err)
 			}
 		})
 	},
 }
-
-// QueryTraceFromTempo queries span count from tempo
-// TODO: move to utils package if needed
-func QueryTraceFromTempo(t *testing.T, c client.Client, tags map[string]string) (int, error) {
-	svc := corev1.Service{}
-	if err := c.Get(context.Background(), types.NamespacedName{
-		Namespace: "monitoring",
-		Name:      "tempo",
-	}, &svc); err != nil {
-		return -1, err
-	}
-	host := ""
-	for _, ing := range svc.Status.LoadBalancer.Ingress {
-		if ing.IP != "" {
-			host = ing.IP
-			break
-		}
-	}
-
-	tagsQueryParam, err := createTagsQueryParam(tags)
-	if err != nil {
-		return -1, err
-	}
-
-	tempoURL := url.URL{
-		Scheme: "http",
-		Host:   net.JoinHostPort(host, "3100"),
-		Path:   "/api/search",
-	}
-	query := tempoURL.Query()
-	query.Add("start", fmt.Sprintf("%d", time.Now().Add(-10*time.Minute).Unix())) // query traces from last 10 minutes
-	query.Add("end", fmt.Sprintf("%d", time.Now().Unix()))
-	query.Add("tags", tagsQueryParam)
-	tempoURL.RawQuery = query.Encode()
-
-	req, err := http.NewRequest("GET", tempoURL.String(), nil)
-	if err != nil {
-		return -1, err
-	}
-
-	t.Logf("send request to %s", tempoURL.String())
-	res, err := http.DefaultClient.Do(req)
-	if err != nil {
-		return -1, err
-	}
-
-	if res.StatusCode != http.StatusOK {
-		return -1, fmt.Errorf("failed to query tempo, url=%s, status=%s", tempoURL.String(), res.Status)
-	}
-
-	tempoResponse := &tempopb.SearchResponse{}
-	if err := jsonpb.Unmarshal(res.Body, tempoResponse); err != nil {
-		return -1, err
-	}
-
-	total := len(tempoResponse.Traces)
-	t.Logf("get response from tempo, url=%s, response=%v, total=%d", tempoURL.String(), tempoResponse, total)
-	return total, nil
-}
-
-// copy from https://github.com/grafana/tempo/blob/c0127c78c368319433c7c67ca8967adbfed2259e/cmd/tempo-query/tempo/plugin.go#L361
-func createTagsQueryParam(tags map[string]string) (string, error) {
-	tagsBuilder := &strings.Builder{}
-	tagsEncoder := logfmt.NewEncoder(tagsBuilder)
-	for k, v := range tags {
-		err := tagsEncoder.EncodeKeyval(k, v)
-		if err != nil {
-			return "", err
-		}
-	}
-	return tagsBuilder.String(), nil
-}

test/e2e/tests/utils.go

@@ -11,6 +11,7 @@ import (
 	"errors"
 	"fmt"
 	"io"
+	"net"
 	"net/http"
 	"net/url"
 	"strings"
@@ -20,7 +21,10 @@ import (
 	"fortio.org/fortio/fhttp"
 	"fortio.org/fortio/periodic"
 	flog "fortio.org/log"
+	"github.com/go-logfmt/logfmt"
+	"github.com/gogo/protobuf/jsonpb" // nolint: depguard // tempopb use gogo/protobuf
 	"github.com/google/go-cmp/cmp"
+	"github.com/grafana/tempo/pkg/tempopb"
 	dto "github.com/prometheus/client_model/go"
 	"github.com/prometheus/common/expfmt"
 	"github.com/stretchr/testify/require"
@@ -494,3 +498,74 @@ type LokiQueryResponse struct {
 		}
 	}
 }
+
+// QueryTraceFromTempo queries span count from tempo
+func QueryTraceFromTempo(t *testing.T, c client.Client, tags map[string]string) (int, error) {
+	svc := corev1.Service{}
+	if err := c.Get(context.Background(), types.NamespacedName{
+		Namespace: "monitoring",
+		Name:      "tempo",
+	}, &svc); err != nil {
+		return -1, err
+	}
+	host := ""
+	for _, ing := range svc.Status.LoadBalancer.Ingress {
+		if ing.IP != "" {
+			host = ing.IP
+			break
+		}
+	}
+
+	tagsQueryParam, err := createTagsQueryParam(tags)
+	if err != nil {
+		return -1, err
+	}
+
+	tempoURL := url.URL{
+		Scheme: "http",
+		Host:   net.JoinHostPort(host, "3100"),
+		Path:   "/api/search",
+	}
+	query := tempoURL.Query()
+	query.Add("start", fmt.Sprintf("%d", time.Now().Add(-10*time.Minute).Unix())) // query traces from last 10 minutes
+	query.Add("end", fmt.Sprintf("%d", time.Now().Unix()))
+	query.Add("tags", tagsQueryParam)
+	tempoURL.RawQuery = query.Encode()
+
+	req, err := http.NewRequest("GET", tempoURL.String(), nil)
+	if err != nil {
+		return -1, err
+	}
+
+	t.Logf("send request to %s", tempoURL.String())
+	res, err := http.DefaultClient.Do(req)
+	if err != nil {
+		return -1, err
+	}
+
+	if res.StatusCode != http.StatusOK {
+		return -1, fmt.Errorf("failed to query tempo, url=%s, status=%s", tempoURL.String(), res.Status)
+	}
+
+	tempoResponse := &tempopb.SearchResponse{}
+	if err := jsonpb.Unmarshal(res.Body, tempoResponse); err != nil {
+		return -1, err
+	}
+
+	total := len(tempoResponse.Traces)
+	t.Logf("get response from tempo, url=%s, response=%v, total=%d", tempoURL.String(), tempoResponse, total)
+	return total, nil
+}
+
+// copy from https://github.com/grafana/tempo/blob/c0127c78c368319433c7c67ca8967adbfed2259e/cmd/tempo-query/tempo/plugin.go#L361
+func createTagsQueryParam(tags map[string]string) (string, error) {
+	tagsBuilder := &strings.Builder{}
+	tagsEncoder := logfmt.NewEncoder(tagsBuilder)
+	for k, v := range tags {
+		err := tagsEncoder.EncodeKeyval(k, v)
+		if err != nil {
+			return "", err
+		}
+	}
+	return tagsBuilder.String(), nil
+}

test/helm/gateway-addons-helm/default.out.yaml

@@ -9059,7 +9059,7 @@ spec:
       dnsPolicy: ClusterFirst
       containers:
         - name: fluent-bit
-          image: "cr.fluentbit.io/fluent/fluent-bit:2.1.4"
+          image: "fluent/fluent-bit:2.1.4"
           imagePullPolicy: Always
           ports:
             - name: http

test/helm/gateway-addons-helm/e2e.out.yaml

@@ -9048,7 +9048,7 @@ spec:
       dnsPolicy: ClusterFirst
       containers:
         - name: fluent-bit
-          image: "cr.fluentbit.io/fluent/fluent-bit:2.1.4"
+          image: "fluent/fluent-bit:2.1.4"
           imagePullPolicy: Always
           ports:
             - name: http

tools/make/kube.mk

@@ -191,6 +191,13 @@ install-e2e-telemetry: helm-generate.gateway-addons-helm
 	helm upgrade -i eg-addons charts/gateway-addons-helm --set grafana.enabled=false,opentelemetry-collector.enabled=true -n monitoring --create-namespace --timeout='$(WAIT_TIMEOUT)' --wait --wait-for-jobs
 	# Change loki service type from ClusterIP to LoadBalancer
 	kubectl patch service loki -n monitoring -p '{"spec": {"type": "LoadBalancer"}}'
+	# Wait service Ready
+	kubectl rollout status --watch --timeout=5m -n monitoring deployment/prometheus
+	kubectl rollout status --watch --timeout=5m statefulset/loki -n monitoring
+	kubectl rollout status --watch --timeout=5m statefulset/tempo -n monitoring
+	# Restart otel-collector to make sure otlp exporter worked
+	kubectl rollout restart -n monitoring deployment/otel-collector
+	kubectl rollout status --watch --timeout=5m -n monitoring deployment/otel-collector
 
 .PHONY: uninstall-e2e-telemetry
 uninstall-e2e-telemetry: