Performance for Cloud Architects
High-performing cloud design means matching resources to demand with low latency. It includes choosing right-sized instance types, caching with CDNs and in-memory stores, picking suitable databases, scaling horizontally with autoscaling and load balancing, and measuring with load tests and monitoring rather than guessing.
Stage: Resilient and Performant · Part of the Cloud Architect roadmap · Last reviewed 19 November 2026
What to learn in Performance
- Place a CDN and cache in front of read-heavy content
- Configure autoscaling from a metric such as CPU or request count
- Choose a database type by access pattern and consistency need
- Run a load test and read latency percentiles
- Find the bottleneck before adding capacity
Performance interview questions
Why look at p95 or p99 latency instead of the average?
Averages hide slow requests. A small share of very slow responses can ruin user experience while the mean looks fine. High percentiles show the tail, which is what many users actually feel and what usually reveals bottlenecks.
What is the difference between vertical and horizontal scaling?
Vertical scaling gives one machine more CPU or memory, which is simple but limited and often needs downtime. Horizontal scaling adds more instances behind a load balancer, which is more resilient but needs stateless design.
Where Performance fits in the Cloud Architect roadmap
Performance belongs to the Resilient and Performant stage (advanced level), which takes about 38 hours. Build for scale and failure.
Other units in this stage
Back to the full Cloud Architect roadmap