The Frugal Architect, Law IV: Unobserved Systems Lead To Unknown Costs

This is part four of seven in our Frugal Architect blog series. Read the rest of the series here.

In case you weren’t as giddy as CloudZero was at re:Invent this past year, we wanted to recount the seven laws outlined by Werner Vogels, Amazon’s CTO, which he’s bundled into a framework called “The Frugal Architect” (check out the whole framework here).

What is “The Frugal Architect”? A constitution of sorts for how engineers can build high-functioning, cost-efficient cloud software. A roadmap to sustainable innovation — and, we have to assume unintentionally on Vogels’s part, the vision by which CloudZero was founded and operates.

In this blog series, we’re going to run through each of the seven laws in “The Frugal Architect,” going over what they mean and why CloudZero is the optimal platform to power each.

Law IV: Unobserved Systems Lead To Unknown Costs

Thus spake Vogels:

Without careful observation and measurement, the true costs of operating a system remain invisible. Like a utility meter tucked away in a basement, lack of visibility enables wasteful habits. Making meters more visible can profoundly shift behaviors.

Though observation requires investment, not implementing adequate monitoring is shortsighted. The adage warns, “If you can’t measure it, you can’t manage it.” Tracking utilization, spending, errors, and more, is crucial for cost management.

When critical cost metrics are placed front and center before engineers and their business partners, more sustainable practices emerge organically. Ongoing inspection lets you spot excess spend and tune operations to trim expenses. The return on investment in observability typically far outweighs the expense.

Most importantly, keeping costs in the forefront encourages sustainable practices.

Let’s break it down.

What engineering teams need: 

… and to invest in cost visibility, whose “return on investment … typically far outweighs the expense.” Vogels gets it.

Why they should do it: Because unobserved systems lead to unknown costs. When you don’t know what a system costs, there’s no way to quantify its business value, justify additional investments, or discourage wasteful habits. Conversely, well-understood costs encourage sustainable engineering practices.

How CloudZero Helps

Track spending, errors, and more on a constant basis

In other words, engineers need relevant, accurate, real-time data on the cost to build and run their cloud infrastructure. This speaks to one of the most important distinctions between platform-aided cost management versus DIY cost management. If you rely solely on CSP-native tools and/or invoices to manage your costs, you put yourself in an inherently reactive position: waiting to get the bill, then trying to optimize in response.

Real-time data, on the other hand, allows you to take a proactive stance: seeing the costs associated with cloud infrastructure as soon as you push the code — and, if the costs look unusual, the power to investigate and make informed efficiency changes.

CloudZero Explorer is the single most sophisticated solution for tracking cloud spend. Fed by CostFormation®, CloudZero’s unparalleled allocation engine, CloudZero Explorer lets you view any costs, through any lens, at any time, at hourly granularity — the industry gold standard.

All sources of cloud spend — be they public cloud, SaaS, PaaS, Kubernetes, or anything else — are unified in a single pane of glass, and all data rises to elite levels of recency and accuracy (just ask the AICPA).

Better yet, CloudZero lets you craft Views so that individual teams can view just the spend they own, and sends anomaly alerts when spend within a particular resource or Dimension looks abnormal. 

Take it from Upstart, who used CloudZero to reduce their cloud spend by $10M: “Once we got [our data] into CloudZero, we looked through all of the custom usage metrics we were getting and pinpointed some runaway spend activities that needed to be resolved,” said Chris Gray, Upstart’s principal engineer. Read the whole case study here.

Put critical cost metrics front and center for engineers and their business partners

Cloud-driven organizations need unifying metrics that quantify the business value of their cloud investments — but getting these metrics is no mean feat. Shared resources, containerization, and the ephemeral nature of the cloud all make it challenging to know where one unit ends and the other begins, much less transform multibillion-line-item AWS bills into clean cost metrics.

In CloudZero parlance, these metrics could take the form of Dimensions and/or unit cost metrics. Examples of common Dimensions include:

Unit cost metrics are custom-built to the unit of value companies are delivering in the cloud. They could include:

One of CloudZero’s strongest differentiators is our ability to give our customers solid unit economics calculations. We partner with you to understand which metric(s) most directly correlate with business value in the cloud. Then, we combine CostFormation and telemetry-driven allocation to assign all resources to the correct units.

Dimensions show up as filterable views in CloudZero Explorer, and unit cost metrics show up as automatically updating dashboards in Analytics. You can configure both to automatically send reports to the appropriate stakeholders at regular intervals, thereby putting critical cost metrics in front of the right people at the right time.

It’s what helped LawnStarter justify the creation of two new engineering teams: “LawnStarter is building two new delivery teams to help us expand into new home care verticals … It’s a direct result of proving efficiency with CloudZero. It’s helping us grow our business the way we want to.”

Keep costs at the forefront of engineering activities

All we’ve described so far is key to the objective of keeping costs at the forefront of engineering activities. But it’s worth spotlighting one last CloudZero feature that helps our customers do this in the most pressing of situations: CloudZero Anomalies.

A consequence of complete cost allocation, in combination with hour-grain data, is that your engineers never miss a beat when it comes to cost irregularities. The data in CloudZero is as accurate and current as technologically possible, and Anomalies, in concert with Views, lets us notify the right engineers when resources they own spike (or dip) in cost.

In one particularly pronounced example, an engineer at a digital fitness brand pushed code that inadvertently changed data retrieval policies for all of their devices from monthly to hourly.

Had the change gone unnoticed, it would have cost the company $3.7M — but CloudZero caught the anomaly when the cost impact was under $30,000. Cost avoidance on that scale has a profound impact on companies’ engineering organizations, budgetary flexibility, and ultimately, business value.

The Joy Of Frugality

Dan Sabath, from Hiya, specifically called out how CloudZero helps his team distinguish “cheap” from “frugal.” “Cheap” means spending as little as possible; “frugal” means maximizing the return on every cloud dollar you invest.

There’s much more to the Frugal Architect framework, and there’s much more we have to say on the subject. Stay tuned for three(!) more exciting new articles in this series.

Author: Dustin Lowman

Dustin Lowman, a FinOps certified practitioner, takes no small pleasure in applying his English major’s sensibility to the SaaS world. Dustin also writes, performs, and records original music, and has never met a cardigan he didn’t like.