Compute in the Cloud

Strip away the marketing names, and a cloud "instance" is a virtual machine running on someone else's hardware - everything you already know about Linux processes, memory, and disk applies directly.

Instance types and sizing

Providers sell VMs in fixed "shapes" - a combination of vCPUs, memory, and sometimes GPU or local disk, bundled into a named size (small, medium, large, and dozens of finer-grained tiers). Picking the right size is mostly about matching your workload's actual CPU and memory profile instead of guessing high "to be safe," which just burns money.

Auto-scaling

An auto-scaling group keeps a fleet of identical instances running, and adjusts the count automatically based on a metric - almost always CPU usage or request count:

if average CPU > 70% for 5 minutes:
    add an instance
if average CPU < 20% for 15 minutes:
    remove an instance

This is the mechanism behind "the site scaled up for the traffic spike and back down afterward" - no human watching a dashboard and deciding in the moment.

Load balancers

A load balancer sits in front of that fleet and spreads incoming traffic across every healthy instance. It also does continuous health checks, and quietly stops sending traffic to any instance that stops responding - so one bad instance doesn't become one bad outage.

Ephemeral vs. persistent

A cloud instance should be treated as disposable - auto-scaling can terminate it at any time, and a crashed instance might simply get replaced rather than repaired. Anything that needs to survive that (a database, uploaded files) belongs on persistent, separately-managed storage - never on the instance's own local disk.

Try it

Every instance is still just a Linux machine you can inspect the normal way:

ps aux

When you're ready, hit Mark complete and move to the next lesson.