Top Performance Challenges in Distributed Architectures
summary
More than a taxonomy this gives concrete failure modes and fixes: an AWS metric fetch that makes 907 API calls, takes 41s and 97 threads until bulk APIs reduce it; an N+1 refactor that turns one quote into 44 product service calls or a search that triggers 26,000 DB calls; a 10MB document flow that causes 69MB network transfer reduced to 31MB; a central encryption service overwhelmed by 316 calls; clear advice on timeouts, jitter and implementing automated regression checks with potometer/Captain in CI.
AI-written summary · about
chapters (8)
- 0:00Start
- 2:23Aws Monitoring
- 5:48N plus 1 Pattern
- 14:54Tight Coupling
- 15:20Breaking the Monolith
- 16:29Inefficient Service Flow
- 23:31Dependencies
- 25:59Recap
related talks
42mArchitecting for Failures in Microservices
34mDistributed Systems - What Can Go Wrong and How to Get It Right
28mAccelerate with confidence through AWS Distributed Load Testing
36mMake It Fast
20mSoftware Performance Optimization and Scaling
25mAvoiding Overload in Distributed Systems
16m"Performance Testing: ..."
40mPerformance Testing and Big Data