Deploying resources in Azure is very easy, so easy that someone who doesn’t even now what is Azure can create a VM by just doing what the portal tells to do. This comes with a very big challenge and that is provisioning a resource with so much power which goes unused and cost big money.
When creating the subscription, the best thing to do is create policies to limit the maximum sku possible for different computes. Either deny creating then or notifying admins when such thing is created.
But what if you have been called in as a consultant to look into the Azure environment and reduce it’s cost.
The most common and the low hanging fruit in most cases are the VM disks.
You can easily cut the cost for Azure infrastructure by many thousands (depending upon the VMs created) by just fine tuning the Disk size or removing the unused ones.
As I said before creating resources is easy and so is deleting them, but with VM the thing is multiple resources are created at once but when deleting you need to make sure you are deleting everything. Many a times I have seen someone delete just the VM and it’s disk remain there orphan.
So the 1st thing to check should be unattached disks.
Once you have deleted the unattached disks (make sure that they are not left there for some reason and are actually orphan). The second thing to check is the used disk space and free space.
Check for all the standard SSD disks. They are the ones whose price double with each size and it’s IOPS remain constant till 8 TiB single disk.
So in such cases you can easily scale down the Disks not using the size to a smaller disk size dropping the price by half. Check the disks pricing here.
If you are able to identify even 20 Disks to scale down from E30 to E20 you’ll be saving ~$10k yearly.
When auditing the cost of the Azure the first thing to check should be the VM disks, these can be scaled up and down easily and could help in saving big bucks if tuned appropriately.