Add base support to deploy gateway collectors

apis/otel/v1alpha1/defaults.go

@@ -397,7 +397,181 @@ service:
 
 	defaultGatewayCPU    = "4"
 	defaultGatewayMemory = "8Gi"
-
+	defaultGatewayConfig = `
+    exporters:
+      sapm:
+        access_token: ${SPLUNK_ACCESS_TOKEN}
+        endpoint: https://ingest.${SPLUNK_REALM}.signalfx.com/v2/trace
+      signalfx:
+        access_token: ${SPLUNK_ACCESS_TOKEN}
+        api_url: https://api.${SPLUNK_REALM}.signalfx.com
+        ingest_url: https://ingest.${SPLUNK_REALM}.signalfx.com
+    extensions:
+      health_check: null
+      http_forwarder:
+        egress:
+          endpoint: https://api.${SPLUNK_REALM}.signalfx.com
+      memory_ballast:
+        size_mib: ${SPLUNK_BALLAST_SIZE_MIB}
+      zpages: null
+    processors:
+      batch: null
+      filter/logs:
+        logs:
+          exclude:
+            match_type: strict
+            resource_attributes:
+            - key: splunk.com/exclude
+              value: "true"
+      k8sattributes:
+        extract:
+          annotations:
+          - from: pod
+            key: splunk.com/sourcetype
+          - from: namespace
+            key: splunk.com/exclude
+            tag_name: splunk.com/exclude
+          - from: pod
+            key: splunk.com/exclude
+            tag_name: splunk.com/exclude
+          - from: namespace
+            key: splunk.com/index
+            tag_name: com.splunk.index
+          - from: pod
+            key: splunk.com/index
+            tag_name: com.splunk.index
+          labels:
+          - key: app
+          metadata:
+          - k8s.namespace.name
+          - k8s.node.name
+          - k8s.pod.name
+          - k8s.pod.uid
+        pod_association:
+        - from: resource_attribute
+          name: k8s.pod.uid
+        - from: resource_attribute
+          name: k8s.pod.ip
+        - from: resource_attribute
+          name: ip
+        - from: connection
+        - from: resource_attribute
+          name: host.name
+      memory_limiter:
+        check_interval: 2s
+        limit_mib: ${SPLUNK_MEMORY_LIMIT_MIB}
+      resource/add_cluster_name:
+        attributes:
+        - action: upsert
+          key: k8s.cluster.name
+          value: ${MY_CLUSTER_NAME}
+      resource/add_collector_k8s:
+        attributes:
+        - action: insert
+          key: k8s.node.name
+          value: ${K8S_NODE_NAME}
+        - action: insert
+          key: k8s.pod.name
+          value: ${K8S_POD_NAME}
+        - action: insert
+          key: k8s.pod.uid
+          value: ${K8S_POD_UID}
+        - action: insert
+          key: k8s.namespace.name
+          value: ${K8S_NAMESPACE}
+      resource/logs:
+        attributes:
+        - action: upsert
+          from_attribute: k8s.pod.annotations.splunk.com/sourcetype
+          key: com.splunk.sourcetype
+        - action: delete
+          key: k8s.pod.annotations.splunk.com/sourcetype
+        - action: delete
+          key: splunk.com/exclude
+      resourcedetection:
+        detectors:
+        - env
+        - system
+        override: true
+        timeout: 10s
+    receivers:
+      jaeger:
+        protocols:
+          grpc:
+            endpoint: 0.0.0.0:14250
+          thrift_http:
+            endpoint: 0.0.0.0:14268
+      otlp:
+        protocols:
+          grpc:
+            endpoint: 0.0.0.0:4317
+          http:
+            endpoint: 0.0.0.0:4318
+      prometheus/collector:
+        config:
+          scrape_configs:
+          - job_name: otel-collector
+            scrape_interval: 10s
+            static_configs:
+            - targets:
+              - ${K8S_POD_IP}:8889
+      signalfx:
+        access_token_passthrough: true
+        endpoint: 0.0.0.0:9943
+      zipkin:
+        endpoint: 0.0.0.0:9411
+    service:
+      extensions:
+      - health_check
+      - memory_ballast
+      - zpages
+      - http_forwarder
+      pipelines:
+        logs/signalfx-events:
+          exporters:
+          - signalfx
+          processors:
+          - memory_limiter
+          - batch
+          receivers:
+          - signalfx
+        metrics:
+          exporters:
+          - signalfx
+          processors:
+          - memory_limiter
+          - batch
+          - resource/add_cluster_name
+          receivers:
+          - otlp
+          - signalfx
+        metrics/collector:
+          exporters:
+          - signalfx
+          processors:
+          - memory_limiter
+          - batch
+          - resource/add_collector_k8s
+          - resourcedetection
+          - resource/add_cluster_name
+          receivers:
+          - prometheus/collector
+        traces:
+          exporters:
+          - sapm
+          processors:
+          - memory_limiter
+          - batch
+          - k8sattributes
+          - resource/add_cluster_name
+          receivers:
+          - otlp
+          - jaeger
+          - zipkin
+      telemetry:
+        metrics:
+          address: 0.0.0.0:8889
+`
 	// the javaagent version is managed by the update-javaagent-version.sh script.
 	defaultJavaAgentVersion = "v1.14.1"
 	defaultJavaAgentImage   = "quay.io/signalfx/splunk-otel-instrumentation-java:" + defaultJavaAgentVersion

apis/otel/v1alpha1/splunkotelagent_types.go

@@ -156,7 +156,7 @@ type AgentSpec struct {
 	// +operator-sdk:csv:customresourcedefinitions:type=spec
 	ClusterReceiver CollectorSpec `json:"clusterReceiver,omitempty"`
 
-	// ClusterReceiver is a Splunk OpenTelemetry Collector deployment used to export data to Splunk APM.
+	// Gateway is a Splunk OpenTelemetry Collector deployment used to export data to Splunk APM.
 	// +kubebuilder:validation:Optional
 	// +operator-sdk:csv:customresourcedefinitions:type=spec
 	Gateway CollectorSpec `json:"gateway,omitempty"`
@@ -187,7 +187,7 @@ type Agent struct {
 	Status AgentStatus `json:"status,omitempty"`
 }
 
-//+kubebuilder:object:root=true
+// +kubebuilder:object:root=true
 
 // AgentList contains a list of SplunkOtelAgent.
 type AgentList struct {

apis/otel/v1alpha1/splunkotelagent_webhook.go

@@ -119,7 +119,7 @@ func (r *Agent) validateCRDAgentSpec() error {
 	spec := r.Spec.Agent
 
 	if spec.Replicas != nil {
-		return fmt.Errorf("`replicas` is not supported by clusterReceiver")
+		return fmt.Errorf("`replicas` is not supported by the agent")
 	}
 
 	return nil
@@ -129,11 +129,11 @@ func (r *Agent) validateCRDClusterReceiverSpec() error {
 	spec := r.Spec.ClusterReceiver
 
 	if spec.Replicas != nil {
-		return fmt.Errorf("`replicas` is not supported by clusterReceiver")
+		return fmt.Errorf("`replicas` is not supported by the clusterReceiver")
 	}
 
 	if spec.HostNetwork {
-		return fmt.Errorf("`hostNetwork` cannot be true for clusterReceiver")
+		return fmt.Errorf("`hostNetwork` cannot be true for the clusterReceiver")
 	}
 
 	return nil
@@ -142,10 +142,6 @@ func (r *Agent) validateCRDClusterReceiverSpec() error {
 func (r *Agent) validateCRDGatewaySpec() error {
 	spec := r.Spec.Gateway
 
-	if r.Spec.Gateway.Enabled != nil && *r.Spec.Gateway.Enabled {
-		return fmt.Errorf("gateway is not supported at the moment")
-	}
-
 	if spec.HostNetwork {
 		return fmt.Errorf("`hostNetwork` cannot be true for clusterReceiver")
 	}
@@ -249,8 +245,6 @@ func (r *Agent) defaultClusterReceiver() {
 
 func (r *Agent) defaultGateway() {
 	spec := &r.Spec.Gateway
-	// TODO(splunk): forcibly disable gateway until we add support for it.
-	spec.Enabled = &[]bool{false}[0]
 	spec.HostNetwork = false
 
 	// The gateway is not enabled by default
@@ -259,8 +253,115 @@ func (r *Agent) defaultGateway() {
 		spec.Enabled = &s
 	}
 
+	if spec.Replicas == nil {
+		s := int32(3)
+		spec.Replicas = &s
+	}
+
+	if spec.Ports == nil {
+		spec.Ports = []v1.ServicePort{
+			{
+				Name:     "otlp",
+				Protocol: "TCP",
+				Port:     4317,
+			},
+			{
+				Name:     "otlp-http",
+				Protocol: "TCP",
+				Port:     4318,
+			},
+			{
+				Protocol: "TCP",
+				Port:     55681,
+			},
+			{
+				Name:     "jaeger-thrift",
+				Protocol: "TCP",
+				Port:     14268,
+			},
+			{
+				Name:     "jaeger-grpc",
+				Protocol: "TCP",
+				Port:     14250,
+			},
+			{
+				Name:     "zipkin",
+				Protocol: "TCP",
+				Port:     9411,
+			},
+			{
+				Name:     "signalfx",
+				Protocol: "TCP",
+				Port:     9943,
+			},
+			{
+				Name:     "http-forwarder",
+				Protocol: "TCP",
+				Port:     6060,
+			},
+		}
+	}
+
+	if spec.Enabled == nil {
+		s := false
+		spec.Enabled = &s
+	}
+
+	if spec.Replicas == nil {
+		s := int32(3)
+		spec.Replicas = &s
+	}
+
+	if spec.Ports == nil {
+		spec.Ports = []v1.ServicePort{
+			{
+				Name:     "otlp",
+				Protocol: "TCP",
+				Port:     4317,
+			},
+			{
+				Name:     "otlp-http",
+				Protocol: "TCP",
+				Port:     4318,
+			},
+			{
+				Protocol: "TCP",
+				Port:     55681,
+			},
+			{
+				Name:     "jaeger-thrift",
+				Protocol: "TCP",
+				Port:     14268,
+			},
+			{
+				Name:     "jaeger-grpc",
+				Protocol: "TCP",
+				Port:     14250,
+			},
+			{
+				Name:     "zipkin",
+				Protocol: "TCP",
+				Port:     9411,
+			},
+			{
+				Name:     "signalfx",
+				Protocol: "TCP",
+				Port:     9943,
+			},
+			{
+				Name:     "http-forwarder",
+				Protocol: "TCP",
+				Port:     6060,
+			},
+		}
+	}
+
 	setDefaultResources(spec, defaultGatewayCPU, defaultGatewayMemory)
 	setDefaultEnvVars(spec, r.Spec.Realm, r.Spec.ClusterName)
+
+	if spec.Config == "" {
+		spec.Config = defaultGatewayConfig
+	}
 }
 
 func setDefaultResources(spec *CollectorSpec, defaultCPU string,

bundle/manifests/otel.splunk.com_agents.yaml

@@ -4222,7 +4222,7 @@ spec:
                     x-kubernetes-list-type: atomic
                 type: object
               gateway:
-                description: ClusterReceiver is a Splunk OpenTelemetry Collector deployment
+                description: Gateway is a Splunk OpenTelemetry Collector deployment
                   used to export data to Splunk APM.
                 properties:
                   args:

bundle/manifests/splunk-otel-collector-operator.clusterserviceversion.yaml

@@ -167,8 +167,8 @@ spec:
           deployment(s).
         displayName: Volumes
         path: clusterReceiver.volumes
-      - description: ClusterReceiver is a Splunk OpenTelemetry Collector deployment
-          used to export data to Splunk APM.
+      - description: Gateway is a Splunk OpenTelemetry Collector deployment used to
+          export data to Splunk APM.
         displayName: Gateway
         path: gateway
       - description: Args is the set of arguments to pass to the OpenTelemetry Collector

config/crd/bases/otel.splunk.com_agents.yaml

@@ -4212,7 +4212,7 @@ spec:
                     x-kubernetes-list-type: atomic
                 type: object
               gateway:
-                description: ClusterReceiver is a Splunk OpenTelemetry Collector deployment
+                description: Gateway is a Splunk OpenTelemetry Collector deployment
                   used to export data to Splunk APM.
                 properties:
                   args:

config/manifests/bases/splunk-otel-collector-operator.clusterserviceversion.yaml

@@ -155,8 +155,8 @@ spec:
           deployment(s).
         displayName: Volumes
         path: clusterReceiver.volumes
-      - description: ClusterReceiver is a Splunk OpenTelemetry Collector deployment
-          used to export data to Splunk APM.
+      - description: Gateway is a Splunk OpenTelemetry Collector deployment used to
+          export data to Splunk APM.
         displayName: Gateway
         path: gateway
       - description: Args is the set of arguments to pass to the OpenTelemetry Collector