feat: Add build reference to the info metrics (#817)

The reference will be from `_PULL_BASE_REF` variable from the cloud
build: https://docs.prow.k8s.io/docs/jobs/

The change also fixes the commit label by using the right variable added
in https://github.com/kubernetes/test-infra/pull/34755/files.

Author: JeffLuoo

Dockerfile

@@ -9,6 +9,7 @@ ENV CGO_ENABLED=0
 ENV GOOS=linux
 ENV GOARCH=amd64
 ARG COMMIT_SHA=unknown
+ARG BUILD_REF
 
 # Dependencies
 WORKDIR /src
@@ -21,7 +22,7 @@ COPY pkg/epp ./pkg/epp
 COPY internal ./internal
 COPY api ./api
 WORKDIR /src/cmd
-RUN go build -ldflags="-X sigs.k8s.io/gateway-api-inference-extension/pkg/epp/metrics.CommitSHA=${COMMIT_SHA}" -o /epp
+RUN go build -ldflags="-X sigs.k8s.io/gateway-api-inference-extension/pkg/epp/metrics.CommitSHA=${COMMIT_SHA} -X sigs.k8s.io/gateway-api-inference-extension/pkg/epp/metrics.BuildRef=${BUILD_REF}" -o /epp
 
 ## Multistage deploy
 FROM ${BASE_IMAGE}

Makefile

@@ -51,10 +51,12 @@ ifdef GO_VERSION
 BUILDER_IMAGE = golang:$(GO_VERSION)
 endif
 
+BUILD_REF ?= $(shell git describe --abbrev=0 2>/dev/null)
 ifdef EXTRA_TAG
 IMAGE_EXTRA_TAG ?= $(IMAGE_REPO):$(EXTRA_TAG)
 SYNCER_IMAGE_EXTRA_TAG ?= $(SYNCER_IMAGE_REPO):$(EXTRA_TAG)
 BBR_IMAGE_EXTRA_TAG ?= $(BBR_IMAGE_REPO):$(EXTRA_TAG)
+BUILD_REF = $(EXTRA_TAG)
 endif
 ifdef IMAGE_EXTRA_TAG
 IMAGE_BUILD_EXTRA_OPTS += -t $(IMAGE_EXTRA_TAG)
@@ -177,6 +179,7 @@ image-build: ## Build the EPP image using Docker Buildx.
 		--build-arg BASE_IMAGE=$(BASE_IMAGE) \
 		--build-arg BUILDER_IMAGE=$(BUILDER_IMAGE) \
 		--build-arg COMMIT_SHA=${GIT_COMMIT_SHA} \
+		--build-arg BUILD_REF=${BUILD_REF} \
 		$(PUSH) \
 		$(LOAD) \
 		$(IMAGE_BUILD_EXTRA_OPTS) ./

cloudbuild.yaml

@@ -12,7 +12,7 @@ steps:
     - GIT_TAG=$_GIT_TAG
     - EXTRA_TAG=$_PULL_BASE_REF
     - DOCKER_BUILDX_CMD=/buildx-entrypoint
-    - GIT_COMMIT_SHA=$COMMIT_SHA
+    - GIT_COMMIT_SHA=$_PULL_BASE_SHA
   - name: gcr.io/k8s-staging-test-infra/gcb-docker-gcloud:v20240718-5ef92b5c36
     entrypoint: make
     args:
@@ -44,5 +44,7 @@ substitutions:
   # _PULL_BASE_REF will contain the ref that was pushed to trigger this build -
   # a branch like 'main' or 'release-0.2', or a tag like 'v0.2'.
   _PULL_BASE_REF: 'main'
+  # _PULL_BASE_SHA will contain the Git SHA of the commit that was pushed to trigger this build.
+  _PULL_BASE_SHA: 'abcdef'
 options:
   substitution_option: ALLOW_LOOSE

pkg/epp/metrics/metrics.go

@@ -36,6 +36,9 @@ const (
 var (
 	// The git hash of the latest commit in the build.
 	CommitSHA string
+
+	// The build ref from the _PULL_BASE_REF from cloud build trigger.
+	BuildRef string
 )
 
 var (
@@ -251,7 +254,7 @@ var (
 			Help:           "General information of the current build of Inference Extension.",
 			StabilityLevel: compbasemetrics.ALPHA,
 		},
-		[]string{"commit"},
+		[]string{"commit", "build_ref"},
 	)
 )
 
@@ -409,7 +412,5 @@ func RecordPrefixCacheMatch(matchedLength, totalLength int) {
 }
 
 func RecordInferenceExtensionInfo() {
-	if CommitSHA != "" {
-		InferenceExtensionInfo.WithLabelValues(CommitSHA).Set(1)
-	}
+	InferenceExtensionInfo.WithLabelValues(CommitSHA, BuildRef).Set(1)
 }

site-src/guides/metrics.md

@@ -34,9 +34,9 @@ curl -i ${IP}:${PORT}/v1/completions -H 'Content-Type: application/json' -d '{
 | inference_model_running_requests                | Gauge     | Number of running requests for each model.             | `model_name`=<model-name>  | ALPHA       |
 | inference_pool_average_kv_cache_utilization  | Gauge            | The average kv cache utilization for an inference server pool.    | `name`=<inference-pool-name>                                                 | ALPHA       |
 | inference_pool_average_queue_size            | Gauge            | The average number of requests pending in the model server queue. | `name`=<inference-pool-name>                                                 | ALPHA       |
-| inference_pool_per_pod_queue_size            | Gauge            | The total number of queue for each model server pod under the inference pool         | `model_server_pod`=<model-server-pod-name> `name`=<inference-pool-name>                             | ALPHA       |
+| inference_pool_per_pod_queue_size            | Gauge            | The total number of queue for each model server pod under the inference pool         | `model_server_pod`=&lt;model-server-pod-name&gt; <br> `name`=&lt;inference-pool-name&gt;                             | ALPHA       |
 | inference_pool_ready_pods                    | Gauge            | The number of ready pods for an inference server pool.            | `name`=&lt;inference-pool-name&gt;                                                 | ALPHA       |
-| inference_extension_info                     | Gauge            | The general information of the current build.                     | `commit`=&lt;hash-of-the-build&gt;                                                 | ALPHA       |
+| inference_extension_info                     | Gauge            | The general information of the current build.                     | `commit`=&lt;hash-of-the-build&gt; <br> `build_ref`=&lt;ref-to-the-build&gt;        | ALPHA       |
 
 
 ## Scrape Metrics