Well-architected cloud is designed for two things: handling demand (scaling) and surviving failure (availability).
Scalability & elasticity
Scalability & Elasticity
Scaling up vs out — and why elasticity makes the cloud efficient.
Tap or hover a part to learn more.
Bigger machine.
Scale up by adding CPU/RAM to an existing instance. Simple, but there's a ceiling and it usually needs downtime — you can only make one box so big.
Check your understanding
1. Adding more instances behind a load balancer is…
2. Elasticity means…
Keep learning
Scale up (bigger instance) has a ceiling; scale out (more instances behind a load balancer) is the cloud-native way. Elasticity adds and removes capacity automatically as demand changes — the key to paying only for what you use.
High availability
High Availability & Redundancy
Keeping services running through failure: load balancing and failover.
Tap or hover a part to learn more.
No single point of failure.
High availability starts by removing single points of failure — duplicate servers, links, power and paths so no one component can take the whole service down.
Check your understanding
1. What does a load balancer do?
2. Roughly how much downtime does 99.9% allow per year?
High availability spreads workloads across multiple instances and availability zones so a single failure doesn't take the service down. Combine elasticity with HA and you get services that are both cost-efficient and resilient.
This is core to cloud operations and every vendor pathway.