Lesson 8 — Core Services

Compute Services

Compute services provide the actual processing power to run applications, ranging from full virtual machines you manage yourself to fully managed container and serverless platforms.

Virtual Machines

A virtual machine instance (like Amazon EC2 or Azure Virtual Machines) gives you a configurable server in the cloud, where you choose the operating system, CPU, and memory, and manage everything above that layer yourself — the most flexible, but also the most hands-on compute option.

Managed Compute Options

Beyond raw virtual machines, providers offer increasingly managed compute options that remove more operational burden: managed container services (running Docker containers without managing the underlying servers) and auto-scaling groups (automatically adding or removing VM instances based on demand).

Common Mistakes

  • Provisioning a virtual machine much larger than actually needed "just in case", wasting money on unused capacity.
  • Not setting up auto-scaling for workloads with variable traffic, leading to either poor performance during spikes or wasted spend during quiet periods.
  • Manually managing servers when a managed compute option would meaningfully reduce operational overhead for the same workload.
  • Forgetting to shut down or deallocate unused virtual machine instances, continuing to incur charges for idle resources.

Professional Tip

Regularly audit running compute instances for ones that are underused or entirely idle — "forgotten" virtual machines left running are one of the most common sources of unnecessary cloud spending.

Your Turn

For a website that experiences a traffic spike every Black Friday, describe how auto-scaling could help manage that variable demand cost-effectively.

Mini Quiz

What does auto-scaling do in a cloud compute environment?