The first of our top tips to reduce Azure waste is actually a double tip, as the first topic discussed is disc storage. Premium disc storage is attached to new VMs by default, and one of the first ways to reduce Azure waste is to opt for a standard storage disc when launching VMs unless a premium storage disc is necessary to add a higher level of security to the data.
When you terminate a VM, the disc storage attached to it doesn’t terminate automatically. This can result in hundreds or thousands of discs being charged for - some at premium rates - until the unattached discs are identified and terminated manually. This tip to reduce Azure waste can contribute significantly to reducing spend in Azure.
Our next top moves on to discuss snapshots. In most cases, businesses are only likely to use a recent snapshot to restore a VM or other asset in the event of a failure. However, Azure accounts are littered with obsolete snapshots that should be deleted but aren’t because the lack of visibility on Azure Cloud makes them difficult to locate.
Identifying and terminating obsolete snapshots is just one of the ways in which businesses can reduce Azure waste by going on a “zombie hunt”. Zombie assets aren’t limited to obsolete snapshots but, also include any assets with very low or zero utilization. Look for idle load balancers, unattached IP addresses, and unused SQL databases to further reduce Azure waste.
While reviewing an Azure account to identify zombie assets, it can also be worthwhile looking for older generation VMs. Although upgrading Azure Classic VMs to Azure Resource Manager VMs doesn’t directly reduce Azure waste, the better functionality and manageability of the upgraded VMs will increase performance and prevent having to launch larger VMs to do the same job.
No conversation about how to reduce Azure waste would be complete without discussing overprovisioned VMs. Allowing too much capacity in a new VM is not unusual, and may be necessary when you don’t know how much demand will be placed on the VM. However, the ongoing monitoring of utilization is necessary to adjust VMs to prevent overprovisioning.
VMs aren’t the only assets that should be constantly monitored. Disc storage can be optimized for capacity, IOPS, and throughput; and - conscious of the tip from above that Microsoft assigns premium storage discs by default - monitoring the utilization of disc storage can help identify unnecessary premium discs, effectively killing two birds with one stone.
Returning to VMs, let’s discuss scheduling start/stop times for non-production assets. Most businesses are aware that using software to schedule start/stop times is more reliable than scheduling scripts and manual processes, but many businesses apply universal “9 to 5” schedules, whereas schedules for when access to the VMs is actually required would be more cost effective.
Finally, let’s talk about storage. Although it’s a generally accepted best practice to reduce Azure waste by moving infrequently accessed data from Hot storage to Cold storage, but many business migrate data to lower cost tiers without considering the cost of retrieval fees and speed of retrieval.