ci: add helm lint values matrix for multi-config coverage (llm-d#540)

Add 6 variant values files in charts/batch-gateway/ci/ to exercise
  template branches that the default values.yaml does not cover:
  globalInferenceGateway, modelGateways, flow control, AIMD disabled,
  image digests, and minimal config.

  Update pre-commit workflow to lint all variants alongside the default.

  Closes llm-d#412

Signed-off-by: Madhu Goutham Reddy Ambati <mambati@redhat.com>
Co-authored-by: Lior Aronovich <243445518+lioraron@users.noreply.github.com>

Author: madhugoutham

.github/workflows/pre-commit.yml

@@ -46,4 +46,8 @@ jobs:
         run: make test-integration
 
       - name: Lint Helm chart
-        run: helm lint charts/batch-gateway
+        run: |
+          helm lint charts/batch-gateway
+          for f in charts/batch-gateway/ci/values-*.yaml; do
+            helm lint charts/batch-gateway -f "$f"
+          done

charts/batch-gateway/ci/values-aimd-disabled.yaml

@@ -0,0 +1,12 @@
+# Exercises: static concurrency path with AIMD disabled.
+# Default has aimd.enabled: true; this tests the false branch.
+global:
+  fileClient:
+    fs:
+      pvcName: "batch-data"
+
+processor:
+  config:
+    concurrency:
+      aimd:
+        enabled: false

charts/batch-gateway/ci/values-flow-control.yaml

@@ -0,0 +1,20 @@
+# Exercises: inferenceObjective conditional (processor-configmap.yaml:47-49),
+# sendFairnessHeader conditional (line 103-104),
+# and OTel env var block (apiserver/processor-deployment.yaml)
+global:
+  fileClient:
+    fs:
+      pvcName: "batch-data"
+  otel:
+    endpoint: "http://otel-collector:4317"
+
+processor:
+  config:
+    globalInferenceGateway:
+      url: "http://inference-gateway:8000"
+      inferenceObjective: "batch-sheddable"
+      requestTimeout: "5m"
+      maxRetries: 3
+      initialBackoff: "1s"
+      maxBackoff: "60s"
+    sendFairnessHeader: true

charts/batch-gateway/ci/values-global-gateway.yaml

@@ -0,0 +1,14 @@
+# Exercises: globalInferenceGateway template path (processor-configmap.yaml:44-70)
+global:
+  fileClient:
+    fs:
+      pvcName: "batch-data"
+
+processor:
+  config:
+    globalInferenceGateway:
+      url: "http://inference-gateway:8000"
+      requestTimeout: "5m"
+      maxRetries: 3
+      initialBackoff: "1s"
+      maxBackoff: "60s"

charts/batch-gateway/ci/values-image-digest.yaml

@@ -0,0 +1,18 @@
+# Exercises: image digest template branches (_helpers.tpl:124-130, 184-190, 239-245).
+# Default uses tag; this tests the digest path (repo@sha256:... instead of repo:tag).
+global:
+  fileClient:
+    fs:
+      pvcName: "batch-data"
+
+apiserver:
+  image:
+    digest: "sha256:abc123"
+
+processor:
+  image:
+    digest: "sha256:def456"
+
+gc:
+  image:
+    digest: "sha256:789ghi"

charts/batch-gateway/ci/values-minimal.yaml

@@ -0,0 +1,6 @@
+# Exercises: default values with only the minimum required override.
+# Validates that all templates render with near-default configuration.
+global:
+  fileClient:
+    fs:
+      pvcName: "batch-data"

charts/batch-gateway/ci/values-model-gateways.yaml

@@ -0,0 +1,23 @@
+# Exercises: modelGateways template path with range loop (processor-configmap.yaml:72-101)
+# Includes inferenceObjective on one model to cover the per-model conditional (line 77-78)
+global:
+  fileClient:
+    fs:
+      pvcName: "batch-data"
+
+processor:
+  config:
+    modelGateways:
+      "llama-3":
+        url: "http://gie-a-epp:8081"
+        inferenceObjective: "batch-sheddable-a"
+        requestTimeout: "5m"
+        maxRetries: 3
+        initialBackoff: "1s"
+        maxBackoff: "60s"
+      "mistral":
+        url: "http://gie-b-epp:8081"
+        requestTimeout: "2m"
+        maxRetries: 1
+        initialBackoff: "1s"
+        maxBackoff: "30s"