Amazon DevOps / SRE Interview Questions

1.Tell me about a time you went above and beyond for an internal customer or service consumer. What extra steps did you take and what was the impact?

2.Describe a time when you had to quickly learn a new technology or tool to solve an urgent production issue. What was your learning approach and how did it change how you tackle similar problems?

3.Describe a time when you saw a critical infrastructure or operational problem that wasn't directly your responsibility, but you stepped in to fix it anyway.

4.Walk me through a time you replaced a complex monitoring or deployment process with something much simpler. What was the original complexity and how did you approach simplifying it?

5.Tell me about an incident response or capacity planning decision you made with limited data during a critical situation. How did you approach the decision-making process?

6.Tell me about a time you helped a junior engineer or team member level up their operational skills. How did you approach the mentoring and what was the outcome?

7.You're on-call for an AWS Lambda function that processes DynamoDB streams, and you notice the error rate spiked from 0.1% to 15% in the last hour. How do you debug this?

8.Code a function that processes CloudWatch alarm notifications from SNS and decides whether to auto-scale an ECS service. Consider cost optimization and response time.

9.We need to migrate a legacy monolith that handles order processing to microservices. The current system processes 50,000 orders per minute during peak. What's your approach?

10.Your ECS service is experiencing memory leaks that cause containers to get OOMKilled every 6 hours. The service handles real-time inventory updates for Amazon.com. How do you solve this?

11.You need to design a deployment pipeline for a service that can't have any downtime and serves 100,000 requests per second. Walk me through your blue-green deployment strategy.

12.Design the monitoring and alerting strategy for a new microservice that will handle Prime membership validations. What metrics matter most and why?

13.Design a backup and disaster recovery system for DynamoDB tables that support Amazon Go stores. The system must handle real-time inventory updates and support point-in-time recovery within 1 hour globally.

14.Design a secrets management system for AWS Lambda functions that need database credentials. The system serves 1 million Lambda invocations per minute across thousands of functions.

15.Design a configuration management system that serves feature flags and configuration to all Amazon.com services. It needs to handle 10 million configuration lookups per second with sub-10ms latency.

16.Design the infrastructure automation that manages EC2 capacity for Prime Video's global streaming during live events like Thursday Night Football. Peak traffic can be 50x normal load.

17.Design a health check system for all AWS services that can determine service health within 30 seconds globally. How do you handle the case where your health check system itself becomes unhealthy?

18.You need to design log aggregation for Alexa voice processing. The system ingests 100TB of logs daily from edge devices globally. How do you make this data searchable for debugging within 5 minutes of generation?

19.Tell me about a time you disagreed with a technical decision your manager or another team was making that would impact system reliability. How did you handle it?

20.Describe a time when you took ownership of fixing a recurring production issue that was affecting multiple teams, even though it wasn't officially assigned to you.

21.Tell me about a time you had to make a quick infrastructure or deployment decision during an outage when you couldn't get approval from your manager. What was your decision-making process?

22.Walk me through a time when you had to convince a development team to change how they deploy or monitor their service because their current approach was creating operational burden. How did you build that relationship and drive the change?

23.Describe a situation where you discovered that a widely-used internal tool or process was fundamentally flawed in a way that could affect multiple services. How did you handle raising this concern and driving a solution?

24.Tell me about a time when you had to deliver a critical infrastructure project under a tight deadline while also maintaining the reliability of existing systems. How did you balance these competing priorities and lead through the pressure?

25.Estimate how much it costs Amazon to serve a single product page view on Amazon.com, including all the infrastructure and dependencies. Walk me through your calculation.

26.AWS Lambda cold start times for a critical API increased from 200ms to 2 seconds after a recent deployment, but only for 15% of invocations. How do you estimate which customers are affected and debug the root cause?

27.Prime Video streaming quality drops 20% globally for 30 minutes during peak evening hours. Nothing was deployed and AWS services show green. How do you estimate the customer impact and what's your investigation approach?

28.You notice that DynamoDB read latency for the shopping cart service increased from 2ms to 8ms over the past month, but there's no single incident. How do you diagnose this gradual degradation?

29.Estimate the compute capacity needed to handle all Alexa voice requests globally during peak usage. Include your assumptions about request volume, processing requirements, and regional distribution.

30.Amazon's recommendation engine response time degrades by 50ms during flash sales, affecting the entire site experience. How do you estimate the revenue impact and what's your mitigation strategy?