CP-37576: Use CloudZero fork of Alloy

app/functions/helmless/default-values.yaml

@@ -304,9 +304,9 @@ components:
     #   per node. This is equivalent to setting defaults.federation.enabled to true.
     #   Uses Prometheus in server mode for metrics collection.
     #
-    # - "clustered": EXPERIMENTAL. Uses Grafana Alloy instead of Prometheus for
-    #   metrics collection. Alloy provides better performance and native
-    #   horizontal scalability.
+    # - "clustered": EXPERIMENTAL. Uses the embedded CloudZero Alloy binary for
+    #   metrics collection instead of Prometheus. Alloy provides better
+    #   performance and native horizontal scalability.
     #
     # - null means automatic mode (currently defaults to "agent").
     mode: null
@@ -406,10 +406,29 @@ components:
     # The agent clustered node component configuration.
     # Only applies when components.agent.mode is set to "clustered".
     clusteredNode:
-      # Container image configuration for Grafana Alloy.
+      # Container image override for the Alloy container.
+      #
+      # By default, the CloudZero Alloy binary is embedded in the agent
+      # image and the Alloy container uses the same image as other agent
+      # containers (components.agent.image). Set repository/tag here only
+      # if you need to override the Alloy container image separately.
       image:
-        repository: docker.io/grafana/alloy
-        tag: v1.11.3
+        repository: null
+        tag: null
+      # Command to run in the Alloy container.
+      #
+      # - null (default): Uses /app/cloudzero-alloy (CloudZero Alloy binary)
+      # - []: Uses the image's default entrypoint (for official Alloy image)
+      # - ["/custom/path"]: Uses the specified command
+      #
+      # To use the official Grafana Alloy image instead of the embedded binary:
+      #
+      #   image:
+      #     repository: docker.io/grafana/alloy
+      #     tag: vX.Y.Z
+      #   command: []
+      #
+      command:
 
       # Resource requirements and limits for the container.
       #

docker/Dockerfile

@@ -5,7 +5,8 @@ ARG DEPLOY_IMAGE=scratch
 # 2. dependencies: Download Go modules (cached per platform)
 # 3. builder: Build Go binaries (cached per platform)
 # 4. certs: Extract certificates from distroless image
-# 5. final: Minimal runtime image with compiled binaries
+# 5. alloy: Extract Alloy binary from CloudZero Alloy image
+# 6. final: Minimal runtime image with compiled binaries
 
 # Stage 1: Base tools installation
 FROM --platform=$BUILDPLATFORM golang:1.25.6-alpine AS base-tools
@@ -69,7 +70,10 @@ RUN --mount=type=cache,target=/go/pkg/mod,id=gomod-$TARGETPLATFORM \
 # Stage 4: Access current certs
 FROM gcr.io/distroless/static-debian12:debug@sha256:20d9c135406d8029d30d59eaaa9d62d2edcd4ec5915dbcda324243c40460e8df AS certs
 
-# Stage 5: Final runtime image
+# Stage 5: Extract Alloy binary from CloudZero Alloy image
+FROM ghcr.io/evan-cz/alloy:latest AS alloy
+
+# Stage 6: Final runtime image
 # Note: For debugging, you can temporarily change the image used for building by
 # passing in something like this to 'docker build':
 #
@@ -113,6 +117,9 @@ COPY --from=builder /go/bin/cloudzero-helmless /app/cloudzero-helmless
 COPY --from=builder /go/bin/cloudzero-scout /app/cloudzero-scout
 COPY --from=builder /go/bin/cloudzero-certifik8s /app/cloudzero-certifik8s
 
+# Copy the Alloy binary from the alloy stage
+COPY --from=alloy /bin/cloudzero-alloy /app/cloudzero-alloy
+
 # Allow the default ENTRYPOINT from busybox to be the default,
 # however run the app as the default command
 CMD ["/app/cloudzero-agent-validator", "-h"]

helm/templates/_helpers.tpl

@@ -838,6 +838,26 @@ configuration will not overwrite values from the first configuration.
 {{- end -}}
 {{- end -}}
 
+{{/*
+Generate container command with special handling:
+- null/not set: Uses provided default command array
+- empty array []: No command output (uses image's default entrypoint)
+- non-empty array: Uses the specified command
+
+Usage: {{ include "cloudzero-agent.generateContainerCommand" (dict "command" .Values.components.agent.clusteredNode.command "default" (list "/app/cloudzero-alloy")) | nindent 10 }}
+*/}}
+{{- define "cloudzero-agent.generateContainerCommand" -}}
+{{- $isEmptyArray := and (kindIs "slice" .command) (empty .command) -}}
+{{- if not $isEmptyArray -}}
+command:
+  {{- if kindIs "invalid" .command }}
+  {{- toYaml .default | nindent 2 }}
+  {{- else }}
+  {{- toYaml .command | nindent 2 }}
+  {{- end }}
+{{- end -}}
+{{- end -}}
+
 {{/*
 Generate image configuration with defaults.
 */}}

helm/templates/agent-deploy.yaml

@@ -227,10 +227,10 @@ spec:
             {{- include "cloudzero-agent.apiKeyVolumeMount" . | nindent 12 }}
         {{- end }}{{/* End Prometheus container */}}
         {{- if eq (include "cloudzero-agent.Values.components.agent.mode" .) "clustered" }}
-        # Grafana Alloy container
+        # CloudZero Alloy container (binary embedded in agent image)
         - name: {{ template "cloudzero-agent.name" . }}-alloy
-          {{- $alloyImage := include "cloudzero-agent.agentCollectorImage" . | fromYaml }}
-          {{- include "cloudzero-agent.generateImage" (dict "defaults" .Values.defaults.image "image" $alloyImage) | nindent 10 }}
+          {{- include "cloudzero-agent.generateImage" (dict "defaults" .Values.defaults.image "image" .Values.components.agent.image "compat" .Values.components.agent.clusteredNode.image) | nindent 10 }}
+          {{- include "cloudzero-agent.generateContainerCommand" (dict "command" .Values.components.agent.clusteredNode.command "default" (list "/app/cloudzero-alloy")) | nindent 10 }}
           args:
             - run
             - /etc/alloy/alloy-config.river

helm/tests/alloy_deployment_test.yaml

@@ -32,7 +32,7 @@ tests:
         template: templates/agent-deploy.yaml
       - matchRegex:
           path: spec.template.spec.containers[1].image
-          pattern: grafana/alloy
+          pattern: cloudzero-agent
         template: templates/agent-deploy.yaml
       - isNull:
           path: data["prometheus.yml"]
@@ -84,6 +84,10 @@ tests:
     set:
       components.agent.mode: clustered
     asserts:
+      - contains:
+          path: spec.template.spec.containers[1].command
+          content: /app/cloudzero-alloy
+        template: templates/agent-deploy.yaml
       - contains:
           path: spec.template.spec.containers[1].args
           content: run

helm/tests/alloy_image_configuration_test.yaml

@@ -2,36 +2,33 @@ suite: test Alloy image configuration
 templates:
   - templates/agent-deploy.yaml
 tests:
-  # Test default Alloy image
-  - it: should use default Alloy image from components.agent.clusteredNode.image
+  # Test that Alloy container uses the agent image (Alloy binary is embedded)
+  - it: should use the agent image for the Alloy container
     set:
       components.agent.mode: clustered
     asserts:
       - matchRegex:
           path: spec.template.spec.containers[1].image
-          pattern: docker.io/grafana/alloy:v1.11.3
+          pattern: cloudzero-agent
 
-  # Test custom Alloy image
-  - it: should use custom Alloy image when specified
+  # Test custom agent image applies to Alloy container
+  - it: should use custom agent image when specified
     set:
       components.agent.mode: clustered
-      components.agent.clusteredNode.image:
-        repository: custom.registry.io/custom/alloy
+      components.agent.image:
+        repository: custom.registry.io/custom/agent
         tag: v2.0.0
     asserts:
       - matchRegex:
           path: spec.template.spec.containers[1].image
-          pattern: custom.registry.io/custom/alloy:v2.0.0
+          pattern: custom.registry.io/custom/agent:v2.0.0
 
-  # Test that alloy image is only used for Alloy
+  # Test that Alloy image config does not affect Prometheus image
   - it: should not affect Prometheus image
     set:
       components.agent.mode: agent
       components.agent.autoscaling:
         enabled: false
-      components.agent.clusteredNode.image:
-        repository: custom.registry.io/custom/alloy
-        tag: v2.0.0
     asserts:
       - matchRegex:
           path: spec.template.spec.containers[1].image
@@ -40,11 +37,11 @@ tests:
           path: spec.template.spec.containers[1].image
           pattern: alloy
 
-  # Test image pull policy from components.agent.clusteredNode.image
-  - it: should respect pullPolicy from components.agent.clusteredNode.image
+  # Test image pull policy from components.agent.image
+  - it: should respect pullPolicy from components.agent.image
     set:
       components.agent.mode: clustered
-      components.agent.clusteredNode.image:
+      components.agent.image:
         pullPolicy: Always
     asserts:
       - equal:

helm/values.schema.json

@@ -5856,6 +5856,16 @@
             "clusteredNode": {
               "additionalProperties": false,
               "properties": {
+                "command": {
+                  "oneOf": [
+                    {
+                      "$ref": "#/$defs/io.k8s.api.core.v1.Container/properties/command"
+                    },
+                    {
+                      "type": "null"
+                    }
+                  ]
+                },
                 "image": {
                   "$ref": "#/$defs/com.cloudzero.agent.image"
                 },

helm/values.schema.yaml

@@ -666,8 +666,19 @@ properties:
             properties:
               image:
                 description: |
-                  Container image configuration for Grafana Alloy.
+                  Container image override for the Alloy container. By default, the Alloy
+                  container uses the agent image (components.agent.image) which embeds the
+                  CloudZero Alloy binary. Set repository/tag here to override separately.
                 $ref: "#/$defs/com.cloudzero.agent.image"
+              command:
+                description: |
+                  Command to run in the Alloy container.
+                  - null (default): Uses /app/cloudzero-alloy (CloudZero Alloy binary)
+                  - [] (empty array): Uses the image's default entrypoint
+                  - ["/custom/path"]: Uses the specified command
+                oneOf:
+                  - $ref: "#/$defs/io.k8s.api.core.v1.Container/properties/command"
+                  - type: "null"
               resources:
                 description: |
                   Resource requirements and limits for the Alloy container.

helm/values.yaml

@@ -304,9 +304,9 @@ components:
     #   per node. This is equivalent to setting defaults.federation.enabled to true.
     #   Uses Prometheus in server mode for metrics collection.
     #
-    # - "clustered": EXPERIMENTAL. Uses Grafana Alloy instead of Prometheus for
-    #   metrics collection. Alloy provides better performance and native
-    #   horizontal scalability.
+    # - "clustered": EXPERIMENTAL. Uses the embedded CloudZero Alloy binary for
+    #   metrics collection instead of Prometheus. Alloy provides better
+    #   performance and native horizontal scalability.
     #
     # - null means automatic mode (currently defaults to "agent").
     mode: null
@@ -406,10 +406,29 @@ components:
     # The agent clustered node component configuration.
     # Only applies when components.agent.mode is set to "clustered".
     clusteredNode:
-      # Container image configuration for Grafana Alloy.
+      # Container image override for the Alloy container.
+      #
+      # By default, the CloudZero Alloy binary is embedded in the agent
+      # image and the Alloy container uses the same image as other agent
+      # containers (components.agent.image). Set repository/tag here only
+      # if you need to override the Alloy container image separately.
       image:
-        repository: docker.io/grafana/alloy
-        tag: v1.11.3
+        repository: null
+        tag: null
+      # Command to run in the Alloy container.
+      #
+      # - null (default): Uses /app/cloudzero-alloy (CloudZero Alloy binary)
+      # - []: Uses the image's default entrypoint (for official Alloy image)
+      # - ["/custom/path"]: Uses the specified command
+      #
+      # To use the official Grafana Alloy image instead of the embedded binary:
+      #
+      #   image:
+      #     repository: docker.io/grafana/alloy
+      #     tag: vX.Y.Z
+      #   command: []
+      #
+      command:
 
       # Resource requirements and limits for the container.
       #

tests/helm/template/alloy.yaml

@@ -966,9 +966,10 @@ data:
         annotations: {}
         autoscaling: null
         clusteredNode:
+          command: null
           image:
-            repository: docker.io/grafana/alloy
-            tag: v1.11.3
+            repository: null
+            tag: null
           resources:
             limits:
               cpu: 100m
@@ -2601,10 +2602,12 @@ spec:
             - name: config-volume
               mountPath: /etc/config
               readOnly: true
-        # Grafana Alloy container
+        # CloudZero Alloy container (binary embedded in agent image)
         - name: cloudzero-agent-alloy
-          image: "docker.io/grafana/alloy:v1.11.3"
+          image: "ghcr.io/cloudzero/cloudzero-agent/cloudzero-agent:1.2.9"
           imagePullPolicy: "IfNotPresent"
+          command:
+            - /app/cloudzero-alloy
           args:
             - run
             - /etc/alloy/alloy-config.river

tests/helm/template/cert-manager.yaml

@@ -883,9 +883,10 @@ data:
         annotations: {}
         autoscaling: null
         clusteredNode:
+          command: null
           image:
-            repository: docker.io/grafana/alloy
-            tag: v1.11.3
+            repository: null
+            tag: null
           resources:
             limits:
               cpu: 100m

tests/helm/template/federated.yaml

@@ -971,9 +971,10 @@ data:
         annotations: {}
         autoscaling: null
         clusteredNode:
+          command: null
           image:
-            repository: docker.io/grafana/alloy
-            tag: v1.11.3
+            repository: null
+            tag: null
           resources:
             limits:
               cpu: 100m

tests/helm/template/istio.yaml

@@ -898,9 +898,10 @@ data:
         annotations: {}
         autoscaling: null
         clusteredNode:
+          command: null
           image:
-            repository: docker.io/grafana/alloy
-            tag: v1.11.3
+            repository: null
+            tag: null
           resources:
             limits:
               cpu: 100m

tests/helm/template/manifest.yaml

@@ -898,9 +898,10 @@ data:
         annotations: {}
         autoscaling: null
         clusteredNode:
+          command: null
           image:
-            repository: docker.io/grafana/alloy
-            tag: v1.11.3
+            repository: null
+            tag: null
           resources:
             limits:
               cpu: 100m