feat: Update evaluation and workflow templates to enhance configuration management and streamline execution steps

Author: fmallmann

deploy/base/templates/evaluate-template.yaml

@@ -12,14 +12,10 @@ spec:
     model:
       type: string
       description: "Model name to use for evaluation (e.g., gemini-2.5-flash-lite)"
-    metricsConfigPath:
+    openApiBasePath:
       type: string
-      description: "Path to metrics configuration file (JSON or YAML)"
-      default: "config/metrics.yaml"
-    image:
-      type: string
-      description: "Docker image to use for the evaluate step"
-      default: "ghcr.io/agentic-layer/testbench/testworkflows:0.1.1"
+      description: "Base path for OpenAI API"
+      default: "http://ai-gateway-litellm.ai-gateway:4000"
 
   # Steps to execute
   steps:
@@ -33,11 +29,14 @@ spec:
           - -c
         args:
           - |
-            uv run python3 evaluate.py "{{ config.model }}" --metrics-config "{{ config.metricsConfigPath }}" && \
+            uv run python3 evaluate.py "{{ config.model }}" --metrics-config "/app/config/metrics.yaml" && \
             if [ -f data/results/evaluation_scores.json ]; then
               echo "✓ Evaluation completed"
               cat data/results/evaluation_scores.json
             else
               echo "✗ Error: Results file not created"
               exit 1
             fi
+        env:
+          - name: OPENAI_API_BASE
+            value: "{{ config.openApiBasePath }}"

deploy/base/templates/publish-template.yaml

@@ -8,6 +8,13 @@ metadata:
     app: testworkflows
 
 spec:
+
+  config:
+    otelExporterOtlpEndpoint:
+      type: string
+      description: "OTel Collector endpoint for exporting traces and metrics"
+      default: "http://lgtm.monitoring:4318"
+
   # Steps to execute
   steps:
     - name: publish-metrics
@@ -17,3 +24,6 @@ spec:
           - "{{ workflow.name }}"
           - "{{ execution.id }}"
           - "{{ execution.number }}"
+        env:
+          - name: OTEL_EXPORTER_OTLP_ENDPOINT
+            value: "{{ config.otelExporterOtlpEndpoint }}"

deploy/base/templates/run-template.yaml

@@ -13,6 +13,10 @@ spec:
     agentUrl:
       type: string
       description: "URL to the agent endpoint (A2A protocol)"
+    otelExporterOtlpEndpoint:
+      type: string
+      description: "OTel Collector endpoint for exporting traces and metrics"
+      default: "http://lgtm.monitoring:4318"
 
   # Steps to execute
   steps:
@@ -24,3 +28,6 @@ spec:
         args:
           - run.py
           - "{{ config.agentUrl }}"
+        env:
+          - name: OTEL_EXPORTER_OTLP_ENDPOINT
+            value: "{{ config.otelExporterOtlpEndpoint }}"

deploy/local/multi-turn-workflow.yaml

@@ -8,57 +8,27 @@ metadata:
     app: testworkflows
 
 spec:
-  # Pod configuration with volumes
-  pod:
-    volumes:
-      - name: metrics-config
-        configMap:
-          name: multi-turn-metrics-config
-      - name: dataset
-        configMap:
-          name: dataset
+  content:
+    files:
+      - path: /data/datasets/ragas_dataset.jsonl
+        contentFrom:
+          configMapKeyRef:
+            name: dataset
+            key: dataset.jsonl
+      - path: /app/config/metrics.yaml
+        contentFrom:
+          configMapKeyRef:
+            name: multi-turn-metrics-config
+            key: metrics.yaml
 
   container:
     image: ghcr.io/agentic-layer/testbench/testworkflows:latest
-    env:
-      - name: OPENAI_API_BASE
-        value: "http://ai-gateway-litellm.ai-gateway:4000"
-      - name: OTEL_EXPORTER_OTLP_ENDPOINT
-        value: "http://lgtm.monitoring:4318"
-    volumeMounts:
-      - name: metrics-config
-        mountPath: /app/config/metrics.yaml
-        subPath: metrics.yaml
-      - name: dataset
-        mountPath: /data/datasets/ragas_dataset.jsonl
-        subPath: dataset.jsonl
 
-  # Steps using the templates
-  steps:
-    # Step 1: Setup - Download and convert dataset
-#    - name: setup
-#      use:
-#        - name: ragas-setup-template
-#          config:
-#            bucket: "datasets"
-#            key: "multi_turn_dataset.json"
-
-    # Step 2: Run - Execute agent queries
-    - name: run
-      use:
-        - name: ragas-run-template
-          config:
-            agentUrl: "http://weather-agent.sample-agents:8000"
-
-    # Step 3: Evaluate - Run RAGAS evaluation
-    - name: evaluate
-      use:
-        - name: ragas-evaluate-template
-          config:
-            model: "gemini-2.5-flash-lite"
-            metricsConfigPath: "/app/config/metrics.yaml"
-
-    # Step 4: Publish - Push metrics to OTLP
-    - name: visualize
-      use:
-        - name: ragas-visualize-template
+  use:
+  - name: ragas-run-template
+    config:
+      agentUrl: "http://weather-agent.sample-agents:8000"
+  - name: ragas-evaluate-template
+    config:
+      model: "gemini-2.5-flash-lite"
+  - name: ragas-visualize-template