Data platform · Platform Concepts

Cost awareness for data platforms

Mediumcostfinopswarehouseincremental

Interview question

How should data engineers think about cloud analytics cost?

How to answer

The short answer

Cloud warehouses and lakes bill for storage, compute scans, and sometimes data transfer. Cost-aware engineering is part of the job.

Levers

  • /cluster so queries scan less
  • Materialize expensive marts instead of recomputing huge views
  • Use / merge instead of full reloads
  • Drop unused tables; set retention on raw zones
  • Avoid SELECT * mega-scans in BI extracts
  • Right-size schedules (do you need every 5 minutes?)
Output
bad:  nightly full reload of 20 TB + ad-hoc SELECT *
good: incremental daily partition + pruned queries

Interview tip: Show you can trade freshness, complexity, and money. Mention measuring cost per job or per dashboard if asked for seniority.