Hello, developers who are diligently coding today! π¨βπ»π©βπ»
Has the company atmosphere subtly changed recently? Has your team leader, who used to shout, “Feature implementation is top priority! Provision plenty of servers!”, started sending you messages like this via Slack DM lately?
“Mr./Ms. OO, the r5.4xlarge instance you launched for testing last week… it’s still running. Is it absolutely necessary right now?” π§
It’s a spine-chilling moment. Yes, it has arrived. The era of free cloud usage is fading, and the grim reaper named ‘FinOps’ has begun knocking on the development team’s door.
Today, we will delve into FinOps in great detail, from a developer’s perspective, even though it might feel like an annoying nag to developers, it’s an unavoidable trend for corporate survival.
1. What Exactly Is FinOps? (Why Is It Tormenting Me?) π€
FinOps is a portmanteau of ‘Finance’ and ‘Operations (DevOps)’. Simply put, it’s ‘a culture and practice for efficiently managing cloud costs’.
Let’s recall the on-premise era. To purchase a server, you had to write a proposal, get approval, and wait months for the equipment to arrive. (CapEx, Capital Expenditure). Spending money was incredibly difficult.
However, the cloud era is different. With a few clicks from a developer, a high-spec server worth millions of won can be created in just one minute. Once a credit card is registered, money is automatically deducted. (OpEx, Operational Expenditure).
The problem arises here.
- The convenience trap: Easy to use, but easy to forget to turn off.
- Complex billing structure: AWS and Azure price lists are like encrypted messages. You don’t even know what you’ve used, but the bill comes with a string of zeros.
- Lack of accountability: When asked, “Who created this?”, everyone claims not to know.
Ultimately, the CFO (Chief Financial Officer), stunned by the ‘bill shock’ that arrived at the end of the month, takes drastic measures.
“From now on, monitor every single won of cloud costs used by the development team!” π€¬πΈ
This is the beginning of FinOps. The finance team has officially started intervening in the engineering domain.
2. FinOps from a Developer’s Perspective: “Nothing Short of a Grim Reaper” π
When FinOps is introduced, the atmosphere in the development team changes significantly. It’s like having a mother-in-law standing behind you while you code, worrying about the electricity bill.
β The Suffocating Tagging Hell π·οΈ
The first step in FinOps is ‘cost allocation’. To bill correctly, you need to know whose server it is and which project it belongs to.
Now, all resources must be mandatorily tagged.
Project: Alpha, Owner: Kim, Environment: Dev… If tags aren’t attached, deployment pipelines can even be blocked. Developers fall into the illusion that they’re spending more time tagging than coding.
β‘ “Is That Absolutely Necessary?” Endless Questioning π΅οΈββοΈ
The finance team or FinOps personnel don’t understand development. They only see numbers.
- “This database’s CPU utilization is 5%, can’t we downgrade its specs?” (But we intentionally provisioned it higher in case of a traffic spike later…)
- “How about turning off development servers on weekends?” (But we might need to access them occasionally on weekends…)
- “Those logs stored in S3 are over a year old, shouldn’t we delete them?” (But we kept them for potential future incident analysis…)
Developers face new stress: having to justify technical decisions to non-developers.
β’ Slowdown in Innovation? π’
Rapid experimentation, like “just launch it and see if it works,” becomes difficult. To test new services, estimated costs must first be reported, and budget approval might be required. Developers begin to complain, “All our time for innovation is being taken away by cost-cutting.”
3. FinOps vs DevOps: Inevitable Culture Clash π₯
This discomfort fundamentally arises because the goals of the two groups are different.
| Category | DevOps (Engineering) π οΈ | FinOps (Finance) π° |
|---|---|---|
| Core Values | Speed, Stability, Performance, Innovation | Predictability, Cost Efficiency, ROI (Return on Investment) |
| Fear | Server crashes (Outages) | Exceeding budget (Cost overruns) |
| Behavioral Pattern | Prepare generously just in case. (Over-provisioning) | Use only what’s strictly necessary. (Rightsizing) |
| Mutual Complaints | “People who don’t understand development only talk about costs.” | “They act like company money is play money.” |
Developers want to provision servers generously out of fear of outages, but FinOps sees all of that as ‘waste’. Bridging this gap is key to FinOps’ success.
4. If You Can’t Avoid It, Embrace It: Survival Strategies in the FinOps Era π
We must acknowledge it. As cloud costs have become one of a company’s largest expenditures, FinOps is an unstoppable trend.
So, what should developers do? Instead of fighting with the finance team, there’s a smart way to respond.
β 1. “Would I Spend My Own Money Like This?” Cultivate Ownership πββοΈ
This is the most basic step. Imagine it’s being paid with your personal credit card, not the company’s.
- Turn off unused development servers when you leave work (or automate it).
- After testing, clean up unattached EBS volumes or unused Elastic IPs. These are hidden cost thieves!
β 2. Cost is Part of ‘Architecture’ (Cost as Architecture) ποΈ
Now, when designing architecture, not only performance and security but also ‘cost’ must be considered as a core factor.
- Instead of “always EC2,” for intermittent tasks, Lambda (serverless) can be much cheaper.
- For batch jobs that can be terminated at any time, propose a design that utilizes Spot Instances to reduce costs by up to 90%. The finance team will praise you.
β 3. Educate the Finance Team (Pre-emptively) π§βπ«
Before the finance team asks, “Why is this so expensive?”, approach them first and explain.
“The architecture we’ve implemented might seem a bit expensive now, but it prevents outages when traffic increases tenfold and ultimately reduces operational labor costs by this much.”
If you explain costs from the perspective of technical debt or risk management, they will understand. FinOps is not one-sided surveillance, but ‘collaboration’.
π Conclusion: FinOps is an Essential Skill for Mature Engineers
At first, FinOps seemed like a grim reaper monitoring your wallet. But upon closer inspection, it’s a rational movement to ‘reduce waste and invest saved money in more valuable areas’.
The era of “developers who only code well” is over. A truly skilled developer is one who can “demonstrate how valuable and efficient their code and infrastructure are from a business perspective”.
Don’t fear the finance team’s Slack messages. It’s just a process of translating the value of your awesome systems into the language of ‘money’. Before you leave work today, why not check if there are any zombie servers you left running and forgot about? ππΈ
Leave a Reply