5 Ways Cloud Costs Can Spiral Out of Control

As CFO of Virtana, I face many of the same challenges as every CFO of a SaaS or enterprise software company today: cost containment, surprises, and an ever-escalating AWS bill. We all need help keeping these things in check. These challenges become even more difficult when your organization goes hybrid cloud. Fortunately, there are tools out there to help our teams help us manage these costs. 

According to the results of our recent State of Hybrid Cloud and FinOps study, only 18% of cloud decision makers surveyed said that they have been able to avoid unnecessary cloud costs. This does not bode well for IT teams responsible for managing hybrid cloud infrastructure within a budget that we CFOs care so much about (i.e., everyone). Where are those excessive costs coming from?

• 41% experienced workloads bursting above agreed capacity
• 35% overprovisioned their compute or storage resources
• 34% over-bought or had unused reserved instances
• 34% stated they had storage blocks no longer attached to a compute instance
• 22% were vexed with poor job scheduling

Half of the respondents reported experiencing two or more of these problems, and 26% were plagued by three or more. Furthermore, given the relative lack of visibility organizations have into their hybrid cloud infrastructures, these numbers are likely understated. Let’s examine each of these budget-busters in turn.

1. Workloads bursting above agreed capacity

Cloud bursting, which is when a workload running on premises or in a private cloud moves or “bursts” into the public cloud when demand spikes, can prevent overload. If you don’t have a good handle on what’s happening with your workloads and how they’re evolving, bursting and the associated costs can escalate well beyond what you’ve budgeted for.

2. Overprovisioning of compute or storage resources

Overprovisioning to support future growth is an age-old approach IT infrastructure teams use on premises. Any costs for capacity you may not be currently using aren’t wasted; these are capex investments that don’t change the amount you pay over the life of those assets. In the cloud, however, because you don’t need to invest in hardware, overprovisioning isn’t necessary, assuming you have an elastic environment. Sometimes it is used to mitigate risk—to ensure that workloads don’t end up under-resourced. The difference is that these “extra” costs are opex that could be eliminated or at least minimized.

3. Over-buying and/or under-utilizing reserved instances

You can save money on your cloud deployment by committing to a particular plan—called a reservation, reserved instance, or programmatic discount—that delivers discounts on the resources you use. While you might have some flexibility to modify or exchange a reserved instance, you typically can’t cancel it, so if you over-commit, the savings might not materialize.

4. Unattached storage blocks

There are a couple of reasons you might have storage blocks that aren’t attached to an active instance. You may need to keep it for audit and compliance purposes, for example. Too often, however, instances get terminated but the attached storage does not—and you continue to pay for that orphaned volume. You need a way to track, report, and eliminate these wasteful, unused resources.

5. Poor job scheduling

If you don’t schedule batch processes or big jobs well, you could end up degrading the performance of other applications. It’s not just a matter of adding capacity. More capacity will always solve the problem, but it might also cost you more than necessary. Instead, you could do things like co-locate a workload that needs to be run at night with another workload that only runs during business hours.

Keeping your cloud costs in check with Virtana Optimize

So, how do we CFOs help our IT/DevOps/Cloud teams keep these cloud costs from spiraling out of control? With Virtana Optimize, you can optimize your cloud capacity and costs on an ongoing basis, so you can stay on budget, even as conditions and options change, and avoid an end-of-month billing surprise. With Optimize you can:

1. Improve forecasting and planning. Use real-time data collection and analytics for ongoing cost management and resource optimization recommendations to meet SLAs and stay on budget.
2. Understand cost-impacting changes in real time. Automatically optimize Amazon EC2 and Azure VM instances with rightsizing recommendations. Quickly adjust to changes in cloud service provider offerings to optimize cost structures.
3. Manage risk more cost-effectively. Tune sizing based on your organization’s risk tolerance with “what if” analysis that includes CPU, memory, I/O, and ingress and egress charges. Safely adjust over-allocated resources to save on your bill without risking production.
4. Optimize reservation planning. Avoid long-term programmatic discount commitment mistakes by finding the ideal resource settings before purchase. Track the amortized value of programmatic discount usage at the instance level.
5. Identify orphaned or unused resources. Automatically identify unused compute instances, storage on stopped instances, and unattached storage blocks. Quickly spot wasted spend with proactively emailed reports.

We need all the help we can get managing our costs in this increasingly complex IT landscape. Fortunately, there are amazing resources available, including our own Virtana Optimize. We have seen our customers truly surprised at the savings they see. The cloud is an amazing resource; now you can keep the costs in check, mitigate surprises, and lower that hosting bill. Amazon may not thank you, but your shareholders will. Request a free trial today!

Get the survey report: State of Hybrid Cloud and FinOps survey

Request a trial: Virtana Optimize