Task

Build a backend service to handle sensor data

Endpoints

• Data Ingestion
  • POST /ingest?url=string (url encoded)
  • Query Parameter
    • url: URL of the CSV file containing sensor data
  • Functionality
    • Add CSV data to the service, retaining older data
    • Synchronous: Retrieval should reflect new data instantly after a successful ingestion
    • Persistent: Data should be retained across restarts and crashes
    • Maximum file size: 5GB
    • Maximum records in CSV: 10M
    • Maximum records in the service: 50M
    • Maximum ingestion time: 60s
• Data Retrieval

  • GET /median?filter=json (url encoded)

  • Query Parameter

    • filter: JSON (optional)
      • id: string[] (optional)
      • type: string[] (optional)
      • subtype: string[] (optional)
      • location: string[] (optional)

  • Response

    • Responses vary depending on the filters applied
    • p95 response time should be less than 10s (max 50 RPS)
    • Median should be calculated on reading column

    { "count": number, "median": number }
    

CSV Structure

Example CSV

https://hack.levels.fyi/data.csv

Submission

• Language: ANY
• Database: ANY
• Submission Format: Docker compose with http service on port 5000
  • Reference: https://docs.docker.com/compose
  • Helper Code: https://gitlab.com/tanishq4/hack (node & python)
  • Maximum containers: 5
• Submission Form: https://forms.gle/skjjEjUz9t1makRW6
  • Only one submission is allowed, subsequent submissions will be ignored
  • The leaderboard may take up to 12 hours to reflect updates
  • Ensure to stress test your service with decent amount of data and requests
  • For questions, please contact [email protected]

Grader

• Grader pulls code from git repository
• Grader runs docker compose up to start your services
• Grader runs multiple ingestions & retrievals test cases (on port 5000)
• Grader runs performance benchmarks & data persistence checks
  • cpus=4 | memory=8GB