Cost Pitfalls of Architecting in the Clouds

Jason Y. Liu
2 min readAug 1, 2021

Building apps is not easy, and many developers understand that its going to be fast and furious — if you don’t build it fast enough, your boss is going to be furious.

It’s a race against time going through waves of iterations and development sprints. Programmers being coders, don’t specialize in getting the infrastructure right.

After all, we just want to write code that works!

While there is nothing wrong with leaving your code and infrastructure to run themselves, having greater visibility and taking an active management approach will pay dividends for years to come. After all, who don’t want to operate reliable, secure, efficient, and cost-effective systems in the cloud?

In this article, I would like to highlight some canaries in the coal mine with regards to Cloud Financial Management, and if any of these resonate with you, don’t hesitate to reach out for both of us to discuss more about the subject of “Cloud Economics”.

Over-provisioned attached managed disks

Engineers provision attached disks as needed, and it is hard to estimate how much disk space we need at the inception.

However, as operations streamline and processes stabilize, there is no need for excess fats in our body.

According to FittedCloud, the average amount of unused storage is about 75%, and we are paying them whether we use them or not. That’s a lot of empty storage that customers are paying for!

Using and spending a ton on commercial software

Overlooked at times, have you realize how much you are spending on virtual servers and how much on licensing, given the same period of time?

I shall not name names, but for a particular OS brand, its almost 1 to 1, between infrastructure costs and licensing costs.

Imagine the amount you can save by going open source!

No idea where costs are going

Cost attribution is a huge genre of discuss as we spend more in the clouds. Which app and which business unit should the costs belong to? Which app is under-performing towards our company bottom line that we should sunset in the foreseeable future?

Plus, throwing in the complexity of container costs that’s running shared services, and Finance and Engineering teams speaking in different languages, it can be as chaotic as an average DOTA match.

Truth to be told, balancing business and technical needs is complex, and I’m still learning every day. Is your organisation in the crawl, walk, or run stage of cloud adoption? Drop me a DM, let’s chat and I want to find out how you are getting your shit together!

--

--