Add Daemonset => Gateway Setup (#4376)

* Add Daemonset => Gateway Setup

* increase limits

* reduce replicas to 3

* reduce batch size to avoid 413

* decrease send_batch_max_size

* disable logs staging

* reset mem

* increase batch size

* reincrease to 1000 for testing

* reset send_batch_size

* remove unnecessary config

.gitlab-ci.yml

@@ -60,7 +60,7 @@ build-ci-image:
     - TAG=ci-image-2
     - docker build --file ci/Dockerfile.gitlab --tag $BUILD_DD_REGISTRY:$TAG .
     - docker push $BUILD_DD_REGISTRY:$TAG
-staging-deploy-gateway-eks:
+.staging-deploy: &staging-deploy
   stage: staging-deploy
   tags: ["runner:docker", "size:large"]
   image: $CI_IMAGE
@@ -79,28 +79,23 @@ staging-deploy-gateway-eks:
     - export AWS_SECRET_ACCESS_KEY=$TEMP_AWS_SECRET_ACCESS_KEY
     # # For debugging
     # - aws sts get-caller-identity
-    - bash ./ci/scripts/ci-deploy-gateway.sh
-
+    - bash $SCRIPT $NAMESPACE $VALUES
 staging-deploy-demo-eks:
-  stage: staging-deploy
-  tags: ["runner:docker", "size:large"]
-  image: $CI_IMAGE
-  dependencies:
-    - push-image-staging
-  rules:
-    - if: '$CI_COMMIT_REF_NAME =~ /-staging$/'
-  script:
-    # # For debugging
-    #- aws sts get-caller-identity
-    - >-
-      TEMP_AWS_ACCESS_KEY_ID=$(aws ssm get-parameter --region us-east-1 --name ci.opentelemetry-collector-contrib.sand-eks-deploy-api-key --with-decryption --query Parameter.Value --out text)
-    - >-
-      TEMP_AWS_SECRET_ACCESS_KEY=$(aws ssm get-parameter --region us-east-1 --name ci.opentelemetry-collector-contrib.sand-eks-deploy-access-key --with-decryption --query Parameter.Value --out text)
-    - export AWS_ACCESS_KEY_ID=$TEMP_AWS_ACCESS_KEY_ID
-    - export AWS_SECRET_ACCESS_KEY=$TEMP_AWS_SECRET_ACCESS_KEY
-    # # For debugging
-    # - aws sts get-caller-identity
-    - bash ./ci/scripts/ci-deploy-staging.sh
+  !!merge <<: *staging-deploy
+  variables:
+    SCRIPT: ./ci/scripts/ci-deploy-staging.sh
+    NAMESPACE: otel-staging
+    VALUES: ./ci/values-staging.yaml
+staging-deploy-gateway-eks:
+  !!merge <<: *staging-deploy
+  variables:
+    SCRIPT: ./ci/scripts/ci-deploy-staging.sh
+    NAMESPACE: otel-gateway
+    VALUES: ./ci/values-gateway.yaml
+staging-deploy-ds-gateway-eks:
+  !!merge <<: *staging-deploy
+  variables:
+    SCRIPT: ./ci/scripts/ci-deploy-ds-gateway.sh
 prod-deploy-demo-eks:
   stage: prod-deploy
   tags: ["runner:docker", "size:large"]

ci/scripts/ci-deploy-gateway.sh → ci/scripts/ci-deploy-ds-gateway.sh

@@ -12,15 +12,24 @@ set -x
 install_collector() {
 	# Set the namespace and release name
 	release_name="opentelemetry-collector"
-	namespace="otel-gateway"
+	release_name_gateway="opentelemetry-collector-gateway"
+	namespace="otel-ds-gateway"
 
 	# if repo already exists, helm 3+ will skip
 	helm --debug repo add open-telemetry https://open-telemetry.github.io/opentelemetry-helm-charts
 
 	# --install will run `helm install` if not already present.
+    # gateway with DD exporter
+	helm --debug upgrade "${release_name_gateway}" -n "${namespace}" open-telemetry/opentelemetry-collector --install \
+		-f ./ci/values.yaml \
+		-f ./ci/values-ds-gateway.yaml \
+		--set-string image.tag="otelcolcontrib-v$CI_COMMIT_SHORT_SHA" \
+		--set-string image.repository="601427279990.dkr.ecr.us-east-1.amazonaws.com/otel-collector-contrib"
+
+    # daemonset with otlp exporter
 	helm --debug upgrade "${release_name}" -n "${namespace}" open-telemetry/opentelemetry-collector --install \
 		-f ./ci/values.yaml \
-		-f ./ci/values-gateway.yaml \
+		-f ./ci/values-otlp-col.yaml \
 		--set-string image.tag="otelcolcontrib-v$CI_COMMIT_SHORT_SHA" \
 		--set-string image.repository="601427279990.dkr.ecr.us-east-1.amazonaws.com/otel-collector-contrib"
 

ci/scripts/ci-deploy-staging.sh

@@ -8,19 +8,19 @@
 set -euo pipefail
 IFS=$'\n\t'
 set -x
+namespace=$NAMESPACE
+values=$VALUES
 
 install_collector() {
-	# Set the namespace and release name
 	release_name="opentelemetry-collector"
-	namespace="otel-staging"
 
 	# if repo already exists, helm 3+ will skip
 	helm --debug repo add open-telemetry https://open-telemetry.github.io/opentelemetry-helm-charts
 
 	# --install will run `helm install` if not already present.
 	helm --debug upgrade "${release_name}" -n "${namespace}" open-telemetry/opentelemetry-collector --install \
 		-f ./ci/values.yaml \
-		-f ./ci/values-staging.yaml \
+		-f "${values}" \
 		--set-string image.tag="otelcolcontrib-v$CI_COMMIT_SHORT_SHA" \
 		--set-string image.repository="601427279990.dkr.ecr.us-east-1.amazonaws.com/otel-collector-contrib"
 

ci/values-ds-gateway.yaml

@@ -0,0 +1,11 @@
+mode: deployment
+replicaCount: 3
+nodeSelector:
+  alpha.eksctl.io/nodegroup-name: ng-6
+clusterRole:
+  name: "opentelemetry-collector-ds-gateway"
+  clusterRoleBinding:
+    name: opentelemetry-collector-ds-gateway
+presets:
+  logsCollection:
+    enabled: false

ci/values-gateway.yaml

@@ -1,4 +1,5 @@
 mode: deployment
+replicaCount: 3
 nodeSelector:
   alpha.eksctl.io/nodegroup-name: ng-5
 clusterRole:

ci/values-otlp-col.yaml

@@ -0,0 +1,25 @@
+nodeSelector:
+  alpha.eksctl.io/nodegroup-name: ng-6
+clusterRole:
+  name: "opentelemetry-collector-ds"
+  clusterRoleBinding:
+    name: opentelemetry-collector-ds
+config:
+  exporters:
+    otlp:
+      endpoint: opentelemetry-collector-gateway:4317
+      tls:
+        insecure: true
+  service:
+    pipelines:
+      metrics:
+        exporters: [otlp]
+      traces:
+        exporters: [otlp]
+      logs:
+        exporters: [otlp]
+resources:
+  limits:
+    # necessary to decrease memory from 2 to 1 Gi in order to
+    # have sufficient memory to schedule all pods on Nodes.
+    memory: 1Gi

ci/values-staging.yaml

@@ -1,2 +1,5 @@
 nodeSelector:
   alpha.eksctl.io/nodegroup-name: ng-3
+presets:
+  logsCollection:
+    enabled: false

ci/values.yaml

@@ -7,7 +7,7 @@ securityContext:
   runAsGroup: 0
 resources:
   limits:
-    cpu: 256m
+    cpu: 512m
     memory: 2Gi
 presets:
   logsCollection: