reenable redis discovery by running the redis server in a dedicated t…

…est (#5782)

* reenable redis discovery by running the redis server in a bridge network

* update discovery

* fix docker host

* run the test in its own test run

* simplify by running the test in isolation

.github/workflows/integration-test.yml

@@ -338,7 +338,7 @@ jobs:
         id: get-matrix
         run: |
           includes=""
-          for service in "apache" "jmx/cassandra" "kafkametrics" "mongodb" "nginx" "envoy" "oracledb" "mysql"; do
+          for service in "apache" "jmx/cassandra" "kafkametrics" "mongodb" "nginx" "envoy" "oracledb" "mysql" "redis"; do
             for arch in "amd64" "arm64"; do
               if [ "$service" = "mongodb" ]; then
                 # tests for mongo "6.0" and "7.0" are flaky, skipping for now

Makefile

@@ -115,11 +115,14 @@ integration-test-envoy-discovery:
 integration-test-nginx-discovery:
 	@set -e; cd tests && $(GOTEST_SERIAL) $(BUILD_INFO_TESTS) --tags=discovery_integration_nginx -v -timeout 5m -count 1 ./...
 
+.PHONY: integration-test-redis-discovery
+integration-test-redis-discovery:
+	@set -e; cd tests && $(GOTEST_SERIAL) $(BUILD_INFO_TESTS) --tags=discovery_integration_redis -v -timeout 5m -count 1 ./...
+
 .PHONY: integration-test-oracledb-discovery
 integration-test-oracledb-discovery:
 	@set -e; cd tests && $(GOTEST_SERIAL) $(BUILD_INFO_TESTS) --tags=discovery_integration_oracledb -v -timeout 5m -count 1 ./...
 
-
 .PHONY: smartagent-integration-test
 smartagent-integration-test:
 	@set -e; cd tests && $(GOTEST_SERIAL) $(BUILD_INFO_TESTS) --tags=smartagent_integration -v -timeout 5m -count 1 ./...

docker/docker-compose.yml

@@ -281,6 +281,7 @@ services:
     image: quay.io/splunko11ytest/redis_server:latest
     profiles:
       - integration
+      - integration-test-redis-discovery
     build: ./redis_server
     ports:
       - "6379:6379"

tests/receivers/redis/bundled_test.go

@@ -12,48 +12,85 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-//go:build integration
+//go:build discovery_integration_redis
 
 package tests
 
 import (
 	"fmt"
-	"runtime"
+	"path/filepath"
 	"testing"
+	"time"
+
+	"github.com/open-telemetry/opentelemetry-collector-contrib/pkg/golden"
+	"github.com/open-telemetry/opentelemetry-collector-contrib/pkg/pdatatest/pmetrictest"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
 
 	"github.com/signalfx/splunk-otel-collector/tests/testutils"
 )
 
 func TestRedisDockerObserver(t *testing.T) {
-	t.Skip("Discovery mode picks up other Docker containers running, making the test fail.")
 	testutils.SkipIfNotContainerTest(t)
-	if runtime.GOOS == "darwin" {
-		t.Skip("unable to share sockets between mac and d4m vm: https://github.com/docker/for-mac/issues/483#issuecomment-758836836")
-	}
-
-	testutils.AssertAllMetricsReceived(t, "all_server_bundled.yaml", "otlp_exporter.yaml",
-		nil, []testutils.CollectorBuilder{
-			func(c testutils.Collector) testutils.Collector {
-				cc := c.(*testutils.CollectorContainer)
-				cc.Container = cc.Container.WithBinds("/var/run/docker.sock:/var/run/docker.sock:ro")
-				cc.Container = cc.Container.WillWaitForLogs("Discovering for next")
-				cc.Container = cc.Container.WithUser(fmt.Sprintf("999:%d", testutils.GetDockerGID(t)))
-				return cc
-			},
-			func(collector testutils.Collector) testutils.Collector {
-				return collector.WithEnv(map[string]string{
-					"REDIS_PASSWORD": "securepassword",
-					"REDIS_USERNAME": "otel",
-					// confirm that debug logging doesn't affect runtime
-					"SPLUNK_DISCOVERY_LOG_LEVEL": "debug",
-				}).WithArgs(
-					"--discovery",
-					"--set", "splunk.discovery.receivers.redis.config.password=${REDIS_PASSWORD}",
-					"--set", "splunk.discovery.receivers.redis.config.username=${REDIS_USERNAME}",
-					"--set", `splunk.discovery.extensions.k8s_observer.enabled=false`,
-					"--set", `splunk.discovery.extensions.host_observer.enabled=false`,
-				)
-			},
-		},
-	)
+	dockerSocket := testutils.CreateDockerSocketProxy(t)
+	require.NoError(t, dockerSocket.Start())
+	t.Cleanup(func() {
+		dockerSocket.Stop()
+	})
+
+	tc := testutils.NewTestcase(t)
+	defer tc.PrintLogsOnFailure()
+	defer tc.ShutdownOTLPReceiverSink()
+
+	_, shutdown := tc.SplunkOtelCollectorContainer("otlp_exporter.yaml", func(c testutils.Collector) testutils.Collector {
+		cc := c.(*testutils.CollectorContainer)
+		cc.Container = cc.Container.WillWaitForLogs("Discovering for next")
+		return cc
+	},
+		func(collector testutils.Collector) testutils.Collector {
+			return collector.WithEnv(map[string]string{
+				"REDIS_PASSWORD": "securepassword",
+				"REDIS_USERNAME": "otel",
+				// confirm that debug logging doesn't affect runtime
+				"SPLUNK_DISCOVERY_LOG_LEVEL": "debug",
+				"SPLUNK_DISCOVERY_DURATION":  "20s",
+			}).WithArgs(
+				"--discovery",
+				"--set", "splunk.discovery.receivers.redis.config.password=${REDIS_PASSWORD}",
+				"--set", "splunk.discovery.receivers.redis.config.username=${REDIS_USERNAME}",
+				"--set", `splunk.discovery.extensions.k8s_observer.enabled=false`,
+				"--set", `splunk.discovery.extensions.host_observer.enabled=false`,
+				"--set", fmt.Sprintf("splunk.discovery.extensions.docker_observer.config.endpoint=tcp://%s", dockerSocket.ContainerEndpoint),
+			)
+		})
+	defer shutdown()
+	expected, err := golden.ReadMetrics(filepath.Join("testdata", "expected.yaml"))
+	require.NoError(t, err)
+	require.EventuallyWithT(t, func(tt *assert.CollectT) {
+		if len(tc.OTLPReceiverSink.AllMetrics()) == 0 {
+			assert.Fail(tt, "No metrics collected")
+			return
+		}
+		err := pmetrictest.CompareMetrics(expected, tc.OTLPReceiverSink.AllMetrics()[len(tc.OTLPReceiverSink.AllMetrics())-1],
+			pmetrictest.IgnoreResourceAttributeValue("service.instance.id"),
+			pmetrictest.IgnoreResourceAttributeValue("net.host.port"),
+			pmetrictest.IgnoreResourceAttributeValue("net.host.name"),
+			pmetrictest.IgnoreResourceAttributeValue("server.address"),
+			pmetrictest.IgnoreResourceAttributeValue("container.name"),
+			pmetrictest.IgnoreResourceAttributeValue("server.port"),
+			pmetrictest.IgnoreResourceAttributeValue("service.name"),
+			pmetrictest.IgnoreResourceAttributeValue("service_instance_id"),
+			pmetrictest.IgnoreResourceAttributeValue("service_version"),
+			pmetrictest.IgnoreMetricAttributeValue("service_version"),
+			pmetrictest.IgnoreMetricAttributeValue("service_instance_id"),
+			pmetrictest.IgnoreTimestamp(),
+			pmetrictest.IgnoreStartTimestamp(),
+			pmetrictest.IgnoreMetricDataPointsOrder(),
+			pmetrictest.IgnoreScopeMetricsOrder(),
+			pmetrictest.IgnoreScopeVersion(),
+			pmetrictest.IgnoreResourceMetricsOrder(),
+			pmetrictest.IgnoreMetricValues(),
+		)
+		assert.NoError(tt, err)
+	}, 60*time.Second, 1*time.Second)
 }