Skip to main content

Troubleshooting

Quick Diagnosis​

# Check container status
docker ps | grep aggregate

# Check recent logs
docker logs aggregate-time-windows --tail 50 2>&1 | grep -i error

# Test with sample data
curl -X POST http://localhost:8080/sensors \
  -H "Content-Type: application/json" \
  -d '{"sensor_id": "temp-001", "temperature": 72, "timestamp": "'$(date -u +%Y-%m-%dT%H:%M:%SZ)'"}'

Common Issues​

Windows not closing / no output​

Cause: Not enough data or timestamp field not recognized

Fix: Check timestamp field name and ensure window period is short enough for testing:

window:
  tumbling:
    size: 1m  # Short for testing
    timestamp_field: timestamp

Aggregates are wrong​

Cause: Data arriving out of order or timestamps in wrong timezone

Fix: Add allowed lateness and normalize timestamps:

window:
  tumbling:
    size: 5m
    allowed_lateness: 30s
- mapping: |
    root.timestamp = this.timestamp.ts_parse("2006-01-02T15:04:05Z07:00").ts_format("2006-01-02T15:04:05Z")

Memory growing over time​

Cause: Window state not being cleaned up

docker stats aggregate-time-windows --no-stream

Fix: Add window expiry and limit group cardinality:

window:
  tumbling:
    size: 5m
    expiry: 1h  # Clean up old state

Late data being dropped​

Cause: allowed_lateness too short

Fix: Increase allowed lateness:

window:
  tumbling:
    size: 5m
    allowed_lateness: 5m  # Increase from default

Still stuck?​

  1. Add debug logging: logger: {level: DEBUG}
  2. Check the Complete Pipeline for reference config
  3. Review Deduplicate Events for handling duplicates