fix(ci): simplify integration tests replay mode (#2997)

We are going to split record and replay workflows completely to simplify
the concurrency key design.

We can add vision tests by just adding to our matrix.

.github/actions/run-and-record-tests/action.yml

@@ -192,7 +192,7 @@ runs:
       if: ${{ always() }}
       uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
       with:
-        name: ${{ inputs.inference-mode }}-logs-${{ github.run_id }}-${{ github.run_attempt || '' }}-${{ inputs.provider }}-${{ inputs.run-vision-tests }}-${{ inputs.stack-config }}
+        name: logs-${{ github.run_id }}-${{ github.run_attempt || '' }}-${{ strategy.job-index }}
         path: |
           *.log
         retention-days: 1

.github/actions/setup-test-environment/action.yml

@@ -16,6 +16,9 @@ inputs:
     description: 'Whether to setup provider for vision tests'
     required: false
     default: 'false'
+  inference-mode:
+    description: 'Inference mode (record or replay)'
+    required: true
 
 runs:
   using: 'composite'
@@ -27,13 +30,13 @@ runs:
         client-version: ${{ inputs.client-version }}
 
     - name: Setup ollama
-      if: ${{ inputs.provider == 'ollama' }}
+      if: ${{ inputs.provider == 'ollama' && inputs.inference-mode == 'record' }}
       uses: ./.github/actions/setup-ollama
       with:
         run-vision-tests: ${{ inputs.run-vision-tests }}
 
     - name: Setup vllm
-      if: ${{ inputs.provider == 'vllm' }}
+      if: ${{ inputs.provider == 'vllm' && inputs.inference-mode == 'record' }}
       uses: ./.github/actions/setup-vllm
 
     - name: Build Llama Stack