Deploy and Manage Azure Compute Resources

What this domain actually tests

The largest domain on the exam by weight: virtual machines and the machinery around them.

Availability sets versus availability zones is the distinction that carries the most marks:

• An availability set spreads VMs across fault domains (different racks, power, network) and update domains (patched at different times) — but all within one datacentre. It protects against rack failure and maintenance reboots.
• Availability zones spread VMs across physically separate datacentres in the region. They protect against a whole datacentre failing.

Zones give the higher SLA. If the requirement is surviving a datacentre outage, a set is not enough.

Then Virtual Machine Scale Sets with autoscaling rules, and App Service plans — where the pricing tier determines what you get (custom domains, SSL, scaling, slots).

The traps

Deployment slots are the zero-downtime answer. An App Service slot is a fully functioning copy of your app at a different endpoint. You deploy to staging, warm it up, verify it, then swap — and the swap is near-instant and reversible. If a question asks how to deploy with no downtime and the ability to roll back immediately, it is a slot swap. Options describing a rolling restart are inferior and are there to catch you.

Disk types matter and are examined. Standard HDD, Standard SSD, Premium SSD, and Ultra Disk. Requirements mentioning consistent low latency or high IOPS eliminate the standard tiers. Also know that the OS disk and data disks are separate and that resizing a VM can change what disk types are available to it.

Scale sets scale, availability sets do not. They are different tools solving different problems, and their similar names invite confusion. A scale set gives you elasticity and fault tolerance; an availability set gives you fault tolerance only.

Resizing a VM usually requires a restart, and some size changes require the VM to be deallocated first (because the target size lives in a different hardware cluster). Questions about changing VM size sometimes hinge on whether downtime is acceptable.

Not every App Service tier supports slots. The free and shared tiers do not. If a scenario needs slots, the plan must be Standard or above.

How to study it

Deploy a VM, then a scale set behind a load balancer, then trigger the autoscale rule by loading it. Watch instances appear.

Then create an App Service with a staging slot, deploy a change to staging, and swap. Then swap back. That five-minute exercise is the answer to a recurring exam question and you will never forget it once you have done it.

Finally, write the sets-versus-zones distinction on a card, with what each protects against. At 25% of the exam this is the biggest domain, but it is also the most conventional — there are fewer nasty surprises here than in networking, and honest hands-on time converts directly into marks.