5 Best Practices for Managing Zombie Cloud Infrastructure

Companies pump up their infrastructures with cloud instances, but after an early flurry of activity they lose track of the resources and, in some cases, forget where things are. They’re left with unused, “zombie” assets. Companies need to adopt strategies to continuously identify and eliminate zombie instances.

In fact, nearly 20 percent of all purchased cloud instances are severely underutilized. This translates to thousands, if not millions, of dollars that would be better used to strengthen security or infrastructure. The machine is on, drawing power and creating bills, but the only measurable output is a usage bill. Companies are unknowingly paying for a surprising amount of infrastructure they don’t end up using.

Cloud service management company CloudHealth Technologies, has offered the following best practices to get this dormant infrastructure back in shape. IT administrators need to identify zombie assets, verify idle status and shut off anything that is unused. Verification is often the hardest part because you must communicate with individuals and teams responsible for running the assets and figure out if there is a business reason to maintain idle assets. These are ongoing conversations that persist as you uncover zombie infrastructure.

Getting Cloud Assets in Shape

Click through for tips on how organizations can get dormant cloud infrastructure back in shape and quit wasting resources on zombie assets, as identified by CloudHealth Technologies.

Identify Zombie Infrastructure

IT must be able to objectively identify when infrastructure is no longer in use. Take a look at C-type EC2 instances – anything that is below 5 percent CPU utilization is likely a zombie. It’s also common to see thousands of dollars of unattached EBS volumes in AWS accounts. The rule of thumb is if a volume has been unattached for two weeks, it’s likely a zombie that can be terminated.

Continuously Right-Size Your infrastructure

Cloud workloads are extremely dynamic by nature – the ability to scale workloads up and down on-demand is one of the biggest benefits of public cloud and can also help to make sure infrastructure doesn’t become dormant. Cloud workloads are also self-service — anyone who needs to provision assets can easily log in and spin them up. Many organizations find compute instances to be extremely underutilized, making them good candidates for rightsizing. There are two main reasons this occurs: either a workload is no longer as resource-intensive as it once was, or someone has accidentally, or purposefully, over-provisioned the instance. Unchecked, this rapidly drives up spend.

Determine the Next Steps for Optimization

Now that we’ve identified potential cases of zombie and underutilized infrastructure, the decision-makers within the IT department need to figure out how to optimize cloud spend and usage that aligns with company budget and personnel. This can be done using reserved instances (RIs) from AWS. RIs can save an organization up to 75 percent compared to on-demand pricing. This is a no-brainer for any company with sustained EC2 or RDS usage.

Flag Cases that Warrant Further Discussion

Organizations of any size will have multiple definitions of what is okay and what needs a second or third look. IT managers need to be on the lookout for potentially troubling underutilization and to constantly be analyzing cloud usage, as we’ve discussed. But further, cloud workload usage needs to be an ongoing conversation within the department to ensure continual success and optimized spend.

Create a Proactive Plan Going Forward

Innovative and forward-thinking managers are the ones that are constantly adapting and adjusting to the needs of the market and their own organization. By establishing a plan for dealing with problems like zombie and underutilized infrastructure in a programmatic fashion, companies can stay ahead of the issue before it costs more time and money.