.github/workflows/monitor.yml

name: Monitor Worker

on:
  schedule:
    - cron: '0 8 * * *' 
  workflow_dispatch:

jobs:
  health-check:
    runs-on: ubuntu-latest
    steps:
      - name: Check worker health
        run: |
          RESPONSE=$(curl -s https://pulsetransit-worker.pablo-matorras.workers.dev/health)
          echo "Worker response: $RESPONSE"
          
          LAST_EST=$(echo $RESPONSE | jq -r '.last_estimaciones')
          LAST_POS=$(echo $RESPONSE | jq -r '.last_posiciones')
          
          # Calculate age
          EST_AGE=$(($(date +%s) - $(date -d "$LAST_EST" +%s)))
          POS_AGE=$(($(date +%s) - $(date -d "$LAST_POS" +%s)))
          
          # Check current hour (UTC - adjust if needed)
          CURRENT_HOUR=$(date +%H)
          
          # Service hours: 5am-11pm UTC (6am-midnight CET)
          if [ $CURRENT_HOUR -ge 5 ] && [ $CURRENT_HOUR -lt 23 ]; then
            # During service hours - strict checks
            if [ $EST_AGE -gt 1800 ]; then  # 30 min
              echo "⛔ ERROR: Estimaciones stale (${EST_AGE}s) during service hours"
              exit 1
            elif [ $EST_AGE -gt 600 ]; then  # 10 min
              echo "⚠️ WARNING: Estimaciones possibly stale (${EST_AGE}s old)"
              # Don't exit - just warn
            fi
            
            if [ $POS_AGE -gt 7200 ]; then
              echo "⛔ ERROR: Posiciones stale (${POS_AGE}s) during service hours"
              exit 1
            fi
          else
            # Outside service hours - lenient
            echo "💤 Outside service hours - data age: est=${EST_AGE}s, pos=${POS_AGE}s"
            if [ $EST_AGE -gt 43200 ]; then  # 12 hours
              echo "⚠️ WARNING: Data very stale, but outside service hours"
            fi
            # Don't fail outside service hours
          fi
          
          echo "✅ Worker is healthy"