Study Azure AZ-305 Compute Fit: key concepts, common traps, and exam decision cues.
Many AZ-305 questions are really asking one architecture question: how much control do you need, and how much operational burden are you willing to keep? Compute choices in Azure usually fall along that line. If you start there, the answer becomes more stable.
| Need | Strongest first fit | Why |
|---|---|---|
| full OS control, custom agents, or legacy dependencies | virtual machines | highest control, highest ops burden |
| web app or API with managed hosting needs | App Service | managed platform with less infrastructure work |
| packaged services needing container orchestration | container platform such as AKS or managed containers | container boundary matters more than VM ownership |
| event-driven short-lived execution | Functions | consumption-style or event-driven compute fit |
The exam often rewards the more managed option unless the scenario clearly requires low-level control.
| Question | If yes, think first about… |
|---|---|
| Do you need OS-level control or custom machine setup? | virtual machines |
| Is the workload mainly a standard web app or API? | App Service |
| Does the team already package and operate containerized services? | container-oriented design |
| Does the workload run in response to events or schedules with variable burst patterns? | Functions or another serverless pattern |
| Trap | Better rule |
|---|---|
| choosing VMs because they feel universal | universal is not the same as best architectural fit |
| choosing Kubernetes because the workload contains containers | container packaging does not always require full cluster orchestration |
| ignoring the operations team skill and support model | AZ-305 designs should match operational reality |
| treating serverless as the answer to every scaling question | execution model and workload shape still matter |