# -----------------------------------------------------------------------------
# router-chaos.yml
# Sprint: SPRINT_5100_0005_0001_router_chaos_suite
# Task: T5 - CI Chaos Workflow
# Description: CI workflow for running router chaos tests.
# -----------------------------------------------------------------------------

name: Router Chaos Tests

on:
  schedule:
    - cron: '0 3 * * *'  # Nightly at 3 AM UTC
  workflow_dispatch:
    inputs:
      spike_multiplier:
        description: 'Load spike multiplier (e.g., 10, 50, 100)'
        default: '10'
        type: choice
        options:
          - '10'
          - '50'
          - '100'
      run_valkey_tests:
        description: 'Run Valkey failure injection tests'
        default: true
        type: boolean

env:
  DOTNET_NOLOGO: 1
  DOTNET_CLI_TELEMETRY_OPTOUT: 1
  TZ: UTC
  ROUTER_URL: http://localhost:8080

jobs:
  load-tests:
    runs-on: ubuntu-22.04
    timeout-minutes: 30

    services:
      postgres:
        image: postgres:16-alpine
        env:
          POSTGRES_USER: stellaops
          POSTGRES_PASSWORD: test
          POSTGRES_DB: stellaops_test
        ports:
          - 5432:5432
        options: >-
          --health-cmd pg_isready
          --health-interval 10s
          --health-timeout 5s
          --health-retries 5

      valkey:
        image: valkey/valkey:7-alpine
        ports:
          - 6379:6379
        options: >-
          --health-cmd "valkey-cli ping"
          --health-interval 10s
          --health-timeout 5s
          --health-retries 5

    steps:
      - name: Checkout
        uses: actions/checkout@v4
        with:
          fetch-depth: 0

      - name: Setup .NET
        uses: actions/setup-dotnet@v4
        with:
          dotnet-version: '10.0.100'
          include-prerelease: true

      - name: Install k6
        run: |
          curl -sSL https://github.com/grafana/k6/releases/download/v0.54.0/k6-v0.54.0-linux-amd64.tar.gz | tar xz
          sudo mv k6-v0.54.0-linux-amd64/k6 /usr/local/bin/
          k6 version

      - name: Cache NuGet packages
        uses: actions/cache@v4
        with:
          path: ~/.nuget/packages
          key: chaos-nuget-${{ runner.os }}-${{ hashFiles('**/*.csproj') }}

      - name: Build Router
        run: |
          dotnet restore src/Router/StellaOps.Router.WebService/StellaOps.Router.WebService.csproj
          dotnet build src/Router/StellaOps.Router.WebService/StellaOps.Router.WebService.csproj -c Release --no-restore

      - name: Start Router
        run: |
          dotnet run --project src/Router/StellaOps.Router.WebService/StellaOps.Router.WebService.csproj -c Release --no-build &
          echo $! > router.pid

          # Wait for router to start
          for i in {1..30}; do
            if curl -s http://localhost:8080/health > /dev/null 2>&1; then
              echo "Router is ready"
              break
            fi
            echo "Waiting for router... ($i/30)"
            sleep 2
          done

      - name: Run k6 spike test
        id: k6
        run: |
          mkdir -p results

          k6 run tests/load/router/spike-test.js \
            -e ROUTER_URL=${{ env.ROUTER_URL }} \
            --out json=results/k6-results.json \
            --summary-export results/k6-summary.json \
            2>&1 | tee results/k6-output.txt

          # Check exit code
          if [ ${PIPESTATUS[0]} -ne 0 ]; then
            echo "k6_status=failed" >> $GITHUB_OUTPUT
          else
            echo "k6_status=passed" >> $GITHUB_OUTPUT
          fi

      - name: Upload k6 results
        if: always()
        uses: actions/upload-artifact@v4
        with:
          name: k6-results-${{ github.run_id }}
          path: results/
          retention-days: 30

      - name: Stop Router
        if: always()
        run: |
          if [ -f router.pid ]; then
            kill $(cat router.pid) 2>/dev/null || true
          fi

  chaos-unit-tests:
    runs-on: ubuntu-22.04
    timeout-minutes: 20
    needs: load-tests
    if: always()

    services:
      postgres:
        image: postgres:16-alpine
        env:
          POSTGRES_USER: stellaops
          POSTGRES_PASSWORD: test
          POSTGRES_DB: stellaops_test
        ports:
          - 5432:5432

      valkey:
        image: valkey/valkey:7-alpine
        ports:
          - 6379:6379

    steps:
      - name: Checkout
        uses: actions/checkout@v4

      - name: Setup .NET
        uses: actions/setup-dotnet@v4
        with:
          dotnet-version: '10.0.100'
          include-prerelease: true

      - name: Build Chaos Tests
        run: |
          dotnet restore tests/chaos/StellaOps.Chaos.Router.Tests/StellaOps.Chaos.Router.Tests.csproj
          dotnet build tests/chaos/StellaOps.Chaos.Router.Tests/StellaOps.Chaos.Router.Tests.csproj -c Release --no-restore

      - name: Start Router for Tests
        run: |
          dotnet run --project src/Router/StellaOps.Router.WebService/StellaOps.Router.WebService.csproj -c Release &
          sleep 15  # Wait for startup

      - name: Run Chaos Unit Tests
        run: |
          dotnet test tests/chaos/StellaOps.Chaos.Router.Tests/StellaOps.Chaos.Router.Tests.csproj \
            -c Release \
            --no-build \
            --logger "trx;LogFileName=chaos-results.trx" \
            --logger "console;verbosity=detailed" \
            --results-directory results \
            -- RunConfiguration.TestSessionTimeout=600000

      - name: Upload Test Results
        if: always()
        uses: actions/upload-artifact@v4
        with:
          name: chaos-test-results-${{ github.run_id }}
          path: results/
          retention-days: 30

  valkey-failure-tests:
    runs-on: ubuntu-22.04
    timeout-minutes: 20
    needs: load-tests
    if: ${{ github.event.inputs.run_valkey_tests != 'false' }}

    steps:
      - name: Checkout
        uses: actions/checkout@v4

      - name: Setup .NET
        uses: actions/setup-dotnet@v4
        with:
          dotnet-version: '10.0.100'
          include-prerelease: true

      - name: Install Docker Compose
        run: |
          sudo apt-get update
          sudo apt-get install -y docker-compose

      - name: Run Valkey Failure Tests
        run: |
          dotnet test tests/chaos/StellaOps.Chaos.Router.Tests/StellaOps.Chaos.Router.Tests.csproj \
            -c Release \
            --filter "Category=Valkey" \
            --logger "trx;LogFileName=valkey-results.trx" \
            --results-directory results \
            -- RunConfiguration.TestSessionTimeout=600000

      - name: Upload Valkey Test Results
        if: always()
        uses: actions/upload-artifact@v4
        with:
          name: valkey-test-results-${{ github.run_id }}
          path: results/

  analyze-results:
    runs-on: ubuntu-22.04
    needs: [load-tests, chaos-unit-tests]
    if: always()

    steps:
      - name: Checkout
        uses: actions/checkout@v4

      - name: Download k6 Results
        uses: actions/download-artifact@v4
        with:
          name: k6-results-${{ github.run_id }}
          path: k6-results/

      - name: Download Chaos Test Results
        uses: actions/download-artifact@v4
        with:
          name: chaos-test-results-${{ github.run_id }}
          path: chaos-results/

      - name: Analyze Results
        id: analysis
        run: |
          mkdir -p analysis

          # Parse k6 summary
          if [ -f k6-results/k6-summary.json ]; then
            echo "=== k6 Test Summary ===" | tee analysis/summary.txt

            # Extract key metrics
            jq -r '.metrics | to_entries[] | "\(.key): \(.value)"' k6-results/k6-summary.json >> analysis/summary.txt 2>/dev/null || true
          fi

          # Check thresholds
          THRESHOLDS_PASSED=true
          if [ -f k6-results/k6-summary.json ]; then
            # Check if any threshold failed
            FAILED_THRESHOLDS=$(jq -r '.thresholds | to_entries[] | select(.value.ok == false) | .key' k6-results/k6-summary.json 2>/dev/null || echo "")

            if [ -n "$FAILED_THRESHOLDS" ]; then
              echo "Failed thresholds: $FAILED_THRESHOLDS"
              THRESHOLDS_PASSED=false
            fi
          fi

          echo "thresholds_passed=$THRESHOLDS_PASSED" >> $GITHUB_OUTPUT

      - name: Upload Analysis
        uses: actions/upload-artifact@v4
        with:
          name: chaos-analysis-${{ github.run_id }}
          path: analysis/

      - name: Create Summary
        run: |
          echo "## Router Chaos Test Results" >> $GITHUB_STEP_SUMMARY
          echo "" >> $GITHUB_STEP_SUMMARY

          echo "### Load Test Results" >> $GITHUB_STEP_SUMMARY
          if [ -f k6-results/k6-summary.json ]; then
            echo "- Total Requests: $(jq -r '.metrics.http_reqs.values.count // "N/A"' k6-results/k6-summary.json)" >> $GITHUB_STEP_SUMMARY
            echo "- Failed Rate: $(jq -r '.metrics.http_req_failed.values.rate // "N/A"' k6-results/k6-summary.json)" >> $GITHUB_STEP_SUMMARY
          else
            echo "- No k6 results found" >> $GITHUB_STEP_SUMMARY
          fi

          echo "" >> $GITHUB_STEP_SUMMARY
          echo "### Thresholds" >> $GITHUB_STEP_SUMMARY
          echo "- Status: ${{ steps.analysis.outputs.thresholds_passed == 'true' && 'PASSED' || 'FAILED' }}" >> $GITHUB_STEP_SUMMARY