Performance / Latency Tail Interview Questions 2026
Published July 22, 2026 · Updated July 22, 2026
13 Performance / Latency Tail interview questions from Amazon, Google, TCS, Apple, Adobe and more. Real questions from Technical, System Design, and HR rounds.
13 questions
Companies: Amazon, Google, TCS, Apple…
All Performance / Latency Tail Questions
Amazon
Hard
System Design round
4-8 years
Design a connection pool for database connections. What parameters would you expose and what are the failure modes?
↑ 49 upvotes · 7 engineers asked this · SDE2
Google
Hard
Scenario round
4-8 years
Your gRPC service has p99 latency of 1 second. Adding a timeout of 500ms seems to help p99 but now 1% of requests fail with DeadlineExceeded. What's the correct approach?
↑ 48 upvotes · 6 engineers asked this · SDE2
Amazon
Hard
System Design round
4-8 years
An engineer proposes adding an in-memory cache to every service to reduce database load. What are the problems with this approach at scale?
↑ 47 upvotes · 5 engineers asked this · SDE2
Amazon
Hard
Technical round
4-8 years
You notice your database CPU is fine but disk I/O is at 100%. Queries are fast individually but slow under concurrent load. What's happening?
↑ 46 upvotes · 4 engineers asked this · SDE2
Amazon
Hard
Technical round
3-6 years
Your API response time is 200ms. You add Redis caching. Response time drops to 20ms for cache hits but is now 400ms for cache misses (double the original). What happened?
↑ 45 upvotes · 3 engineers asked this · SDE2
TCS
Hard
Scenario round
3-6 years
You're building a feature that requires reading from a database 1000 times per request (for each item in a list). Batching is not possible due to the library you must use. What do you do?
↑ 40 upvotes · 28 engineers asked this · SDE2
Apple
Hard
Technical round
3-7 yearsPerformanceiOS
An iOS app's scroll performance drops to visibly janky frame rates on a list with complex cells. Walk through your process for diagnosing whether the bottleneck is layout, rendering, or data processing.
↑ 29 upvotes · 10 engineers asked this · Senior SDE
Adobe
Medium
Technical round
2-6 yearsPerformanceJavaScript
A web-based PDF viewer becomes unresponsive when rendering a 200-page document. What rendering strategy would you use so the app stays responsive?
↑ 21 upvotes · 7 engineers asked this · Senior SDE
Apple
Medium
Technical round
3-7 yearsPerformance
An app's cold launch time regressed from 800ms to 1.4s after a release. What's your systematic approach to finding the regression before assuming which commit caused it?
↑ 21 upvotes · 10 engineers asked this · Senior SDE
Amazon
Medium
Technical round
2-5 years
A microservice that was handling 500 req/sec now crashes at 100 req/sec after adding a new feature. The feature adds one line: logging the full request body. What's happening?
↑ 10 upvotes · 38 engineers asked this · SDE2
Amazon
Hard
System Design round
4-7 years
Your team wants to add caching everywhere to improve performance. What questions would you ask before adding any cache?
↑ 9 upvotes · 37 engineers asked this · SDE2
Amazon
Hard
Scenario round
3-7 years
You're asked to reduce your service's AWS costs by 40% without degrading performance. Where do you start?
↑ 8 upvotes · 36 engineers asked this · SDE2
Amazon
Hard
Scenario round
4-8 years
Your API's p99 latency is 3 seconds but p50 is 50ms. Most users have a great experience but 1% see extreme slowness. How do you investigate the outliers?
↑ 7 upvotes · 35 engineers asked this · SDE2
Practice these questions with AI feedback
Get instant grading on your answers, identify your weak areas, and generate a personalised 14-day study plan — all free.
Build my study plan →
Same topic at specific companies
Related topics
Frequently asked questions
Design a connection pool for database connections. What parameters would you expose and what are the failure modes?
Design a connection pool for database connections. What parameters would you expose and what are the failure modes?
Your gRPC service has p99 latency of 1 second. Adding a timeout of 500ms seems to help p99 but now 1% of requests fail with DeadlineExceeded. What's the correct approach?
Your gRPC service has p99 latency of 1 second. Adding a timeout of 500ms seems to help p99 but now 1% of requests fail with DeadlineExceeded. What's the correct approach?
An engineer proposes adding an in-memory cache to every service to reduce database load. What are the problems with this approach at scale?
An engineer proposes adding an in-memory cache to every service to reduce database load. What are the problems with this approach at scale?
You notice your database CPU is fine but disk I/O is at 100%. Queries are fast individually but slow under concurrent load. What's happening?
You notice your database CPU is fine but disk I/O is at 100%. Queries are fast individually but slow under concurrent load. What's happening?
Your API response time is 200ms. You add Redis caching. Response time drops to 20ms for cache hits but is now 400ms for cache misses (double the original). What happened?
Your API response time is 200ms. You add Redis caching. Response time drops to 20ms for cache hits but is now 400ms for cache misses (double the original). What happened?