Add E2E test matrix parallelization (Phase 1 tests + categorization)

.github/actions/run-e2e-tests/action.yml

@@ -3,8 +3,12 @@ description: Run E2E tests for a specific vertical with Selenium/Chrome
 
 inputs:
   vertical:
-    description: 'Vertical name (insurance/retail/landing)'
+    description: 'Vertical name (insurance/retail/healthcare/landing)'
     required: true
+  category:
+    description: 'Test category (smoke/api/ui/lifecycle) - runs all tests if not specified'
+    required: false
+    default: ''
   web-url:
     description: 'Web URL for the vertical'
     required: true
@@ -47,4 +51,14 @@ runs:
         HEADLESS: 'true'
       run: |
         cd tests/e2e
-        pytest -v --tb=short -m "${{ inputs.vertical }}"
+
+        # Build marker expression based on inputs
+        MARKER="${{ inputs.vertical }}"
+
+        # Add category filter if specified
+        if [ -n "${{ inputs.category }}" ]; then
+          MARKER="${MARKER} and ${{ inputs.category }}"
+        fi
+
+        echo "Running tests with marker: ${MARKER}"
+        pytest -v --tb=short -m "${MARKER}"

tests/e2e/TEST-MATRIX-STRATEGY.md

@@ -0,0 +1,122 @@
+# E2E Test Matrix Parallelization Strategy
+
+## Overview
+Split E2E tests into categories (smoke/api/ui/lifecycle) and run them in parallel per vertical using GitHub Actions matrix strategy.
+
+## Test Categories
+
+### Smoke Tests (fast, ~30-60s per vertical)
+- Page loads, navigation exists, basic rendering
+- No API calls, minimal interactions
+- Run on every PR
+
+### API Tests (~2-3min per vertical)
+- Create/read/delete via API
+- Backend integration, data persistence
+- Run on every PR
+
+### UI Tests (~3-4min per vertical)
+- Interactive workflows: buttons, forms, modals
+- Customer portals, chatbots, data seeding
+- Run on every PR
+
+### Lifecycle Tests (~5-7min per vertical)
+- Complete user journeys (quote→policy→claim)
+- Cross-entity workflows
+- Run on main branch merges (optional for PRs)
+
+## Matrix Configuration
+
+```yaml
+strategy:
+  fail-fast: false
+  matrix:
+    include:
+      # Insurance (12 tests total: 2 smoke, 3 api, 4 ui, 1 lifecycle)
+      - vertical: insurance
+        category: smoke
+        deploy_job: deploy-insurance-ui
+      - vertical: insurance
+        category: api
+        deploy_job: deploy-insurance-ui
+      - vertical: insurance
+        category: ui
+        deploy_job: deploy-insurance-ui
+      - vertical: insurance
+        category: lifecycle
+        deploy_job: deploy-insurance-ui
+
+      # Retail (13 tests total: 5 smoke, 5 api, 2 ui, 1 lifecycle)
+      - vertical: retail
+        category: smoke
+        deploy_job: deploy-retail-ui
+      - vertical: retail
+        category: api
+        deploy_job: deploy-retail-ui
+      - vertical: retail
+        category: ui
+        deploy_job: deploy-retail-ui
+      - vertical: retail
+        category: lifecycle
+        deploy_job: deploy-retail-ui
+
+      # Healthcare (11 tests total: 5 smoke, 4 api, 1 ui, 1 lifecycle)
+      - vertical: healthcare
+        category: smoke
+        deploy_job: deploy-healthcare-ui
+      - vertical: healthcare
+        category: api
+        deploy_job: deploy-healthcare-ui
+      - vertical: healthcare
+        category: ui
+        deploy_job: deploy-healthcare-ui
+      - vertical: healthcare
+        category: lifecycle
+        deploy_job: deploy-healthcare-ui
+
+      # Landing (7 tests total: 4 smoke, 3 ui)
+      - vertical: landing
+        category: smoke
+        deploy_job: deploy-landing-ui
+      - vertical: landing
+        category: ui
+        deploy_job: deploy-landing-ui
+```
+
+## Benefits
+
+1. **Parallelization**: 14 jobs run concurrently instead of 4 sequential
+2. **Fast feedback**: Smoke tests (4 jobs) complete in ~1 min
+3. **Targeted reruns**: Failed API tests don't require rerunning UI tests
+4. **Resource efficiency**: Lifecycle tests can be main-only
+5. **Clear organization**: Failures immediately show which category broke
+
+## Timing Estimates
+
+### PR Workflow (parallel execution):
+- Smoke tests: **1 minute** (4 jobs × 30-60s)
+- API tests: **3 minutes** (3 jobs × 2-3min)
+- UI tests: **4 minutes** (4 jobs × 3-4min)
+- **Total PR time: ~4 minutes** (vs 16 minutes sequential)
+
+### Main Branch (adds lifecycle):
+- Lifecycle tests: **7 minutes** (3 jobs × 5-7min)
+- **Total main time: ~7 minutes** (parallel)
+
+## Test Count Summary
+
+- **Insurance**: 8 → 12 tests (+4 Phase 1)
+- **Retail**: 12 → 13 tests (+1 Phase 1)
+- **Healthcare**: 6 → 11 tests (+5 Phase 1)
+- **Landing**: 7 tests (unchanged)
+- **Total**: 33 → 43 tests (+10 Phase 1)
+
+## Implementation
+
+1. ✅ Add api/ui/lifecycle markers to pytest.ini
+2. ✅ Tag all existing tests with category markers
+3. ✅ Write Phase 1 new tests (10 tests)
+4. ✅ Update run-e2e-tests action to support category parameter
+5. 🔄 Replace individual test jobs with matrix in deploy-test.yml
+6. 🔄 Replace individual test jobs with matrix in deploy-production.yml
+7. ⏳ Verify all tests pass in new structure

tests/e2e/pytest.ini

@@ -19,6 +19,9 @@ markers =
     customer: Customer portal tests
     slow: Slow tests
     vertical: Multi-vertical architecture tests
+    api: API integration tests
+    ui: UI interaction tests
+    lifecycle: Full lifecycle workflow tests
 
 # Test discovery
 python_files = test_*.py