MidCost

Cloud spend jumped 40% this month with no traffic increase. How do you find out why?

What they are really testing: Whether you can investigate cost like an engineer rather than asking teams to be careful. They want a method that gets to a specific line item.

A real interview question

Cloud spend jumped 40% this month with no traffic increase. How do you find out why?

What most people say

drag me

I would ask teams to review their resources and shut down anything unused.

It delegates the investigation instead of doing it, and it will surface a few idle instances while missing the actual cause. A 40% jump usually has one specific driver that a bill diff finds in under an hour.

The follow-ups they ask next

  • Data transfer is the top line. What specifically causes that?

    Cross-AZ chatter between services that used to be co-located, egress to the internet that could use a private endpoint, or NAT gateway processing charges for traffic that should use a VPC endpoint.

  • How do you stop this being a surprise next month?

    Anomaly detection on daily spend rather than a monthly total, cost visible per team through tagging, and a plan or cost estimate shown on infrastructure pull requests.

What the interviewer is listening for

  • Diffs the bill methodically
  • Separates rate from usage
  • Correlates against change history

What sinks the answer

  • Delegates instead of investigating
  • Only suggests reserved instances
  • No prevention story

If you genuinely do not know

Say this instead of freezing. Reasoning out loud from what you do know beats silence every single time, and a good interviewer is listening for exactly that.

I [diff the bill month over month by service, then by resource], and ask [is it rate or usage]. Common flat-traffic causes are [orphans, unattached volumes, cross-AZ or NAT data transfer, log volume]. I [correlate the curve with deploy history], then add [tagging, anomaly alerts, and an orphan sweep].

Keep going with cost

All 87 devops engineer questions

Knowing the answer is not the same as recalling it under pressure

Sign in to send the questions you fumble to spaced recall, so they come back right before you would forget them, and learn the concepts behind them with hands-on labs.

Start free