Quick Win #1

Idle and Unattached

Quick Win #1

Turn Off and/or Delete All Your Zombies and Orphans

Note: For help on this quick win compilation, we spoke to Stratascale internal experts. We also asked several cloud cost management vendors for their take. Good news! Most tended to agree on the top wasters of cloud spend. Several also maintained that such quick wins are great but may be short-lived unless they are also accompanied by other, long-term practices toward better governance and better cloud cost management.

Now let’s look at our first quick win …

The cloud’s biggest benefit is its elasticity. Yet, for controlling costs, cloud’s on-demand provisioning poses challenges. It’s just too easy to keep running (and paying for) a bunch of zombie or orphan cloud resources. (Don’t know what a zombie or orphan is in the land of cloud? See later in this report.)

When we talk about zombies and orphans in the cloud, we’re really talking about things like:

Cloud resources tied to past development sandbox experiments
Virtual servers now sitting idle
Processes and/or storage volumes or persistent disks once attached to something, but now detached and floating around without a parent in sight
Snapshots to disks that no longer exist (or older snapshots that are no longer needed)
Unused external IP addresses that may have once been attached to (or reserved for) a virtual instance

Did you know such now-detached and unused volumes, servers, or ancient logs/snapshots contribute a decent chunk to your cloud bill? According to Turbonomic’s ParkMyCloud blog, about “$2.6 billion” of cloud spending was likely wasted in 2020 on “orphaned volumes and snapshots”

Unfortunately, many organizations may not even be aware of the orphans and zombies they currently have eating up valuable budget.

Why the Fuss about Zombies and Orphans?

Snow Software’s Kim King, director of product marketing for hybrid cloud management, says “one of the biggest culprits of cloud overspend is neglecting to terminate cloud instances when they are no longer needed.” A CloudHealth by VMware report, reiterates these sentiments for reducing multicloud spend. Four of the report’s nine best practices cite deleting or terminating unused resources.

Stratascale Field CTO Jason Hood has other good advice about the developer’s sandbox. “Be on the lookout for ‘sandbox costs’ too, which can happen when a business stands up new environments to support app development and then forgets to shut them down.”

The Duckbill Group’s Quinn cuts to the chase with this cost-cutting guidance: “Turn it off or delete it, full stop,” he says.

“You don’t need hourly snapshots of easily recreated web servers, you don’t need your database server logs from 2012, and you absolutely don’t need that development cluster spun up for an experiment two years ago that needed another 200 terabytes of test data, run by a developer who quit six months ago. You don’t pay for what you use in the cloud; you pay for what you forgot to turn off. So, go turn it off.”

(Do you know the different cloud resources that might be good candidates for deletion? Ask us for a more comprehensive cheat sheet to help get you started.)

Hunt for Zombies

A CloudHealth report advises the following: Start looking for zombies by trying to find compute services from the last 30 days with a max CPU of less than 5%. They may not all be zombies, but it’s a good place to start.

A Closer Look

Zombies and Orphans

Don’t know what a zombie or orphan is when it comes to cloud? See these definitions.

An older IBM blog warns of the danger of zombie cloud resources:

“Cloud zombies are unwanted or forgotten-about cloud services that no one uses, but that continue to be active indefinitely. They are indeed the living dead. And they will continue to feast on your credit cards and billing accounts until you kill them off!”

(All jokes aside, zombie cloud resources also pose a security risk — an excellent reason to hunt them down and kill them without further delay. Zombies, in terms of security, can also have a more sinister definition: a compromised process or service operated by threat actors via botnet, also known as zombie bots.)

Orphaned resources are sadder yet. Turbonomic’s ParkMyCloud blog offers a good definition. It says:

“A resource can become ‘orphaned’ when it is detached from the infrastructure it was created to support, such as a volume detached from an instance or a snapshot detached from any volumes. Whether you are aware these remain in your cloud environment or not, they can continue to incur costs, wasting money and driving up your cloud bill.”