A few months ago, in our organisation, we encountered a situation where we saw an unexpected spike in our AWS bill at the end of the month. This prompted a thorough investigation, and we realized that we hadn't accounted for a particular scenario at production scale, leading to an unexpected overspend. This experience initiated a mini-project within the organization focused on implementing alerting measures to avoid such billing surprises. I’m excited to say that the small mistake we made earlier has now helped us avoid a 3x monthly bill due to an application misconfiguration! I’d like to share some really simple yet powerful approaches to saving your organization from cost surprises.
AWS offers powerful built-in tools and mechanisms for cloud cost management, making it easier to handle Cloud FinOps. Here are the top 3 ways to go about it:
- Configure daily billing alerts
- Set up cost anomaly detection
- Use Cost Explorer for root cause detection
Configure daily billing alerts
Alerts help you respond to events quicker, and there are two ways to create billing alerts in AWS:
a. Create a CloudWatch billing alarm
b. Create an AWS cost budget
You can use both depending on your needs, though the cost budget approach is my personal favorite for its ease of configuration and rich features. Create a budget with the following parameters:
- Period: Daily
- Budget amount: Your expected daily spending
- Scope: 'All AWS Services'
Configure three alert thresholds:
- At the lowest point possible (e.g., 80%) – This ensures you receive the previous day's spending email daily, regardless of whether the budget is exhausted.
- At 100% – To alert you when your budget is hit.
- At a higher conservative point (e.g., 125%) – To notify you when your ultimate limit is reached, signaling that it’s time to take action.
Make sure to revise your budget and thresholds as your AWS spending evolves.
Set up cost anomaly detection
Using machine intelligence to detect unusual cost patterns across AWS services and accounts can make life easier when it comes to spotting and analyzing cost spikes. AWS offers a free tool (free as of the time of writing this blog) called Cost Anomaly Detection, which allows you to create cost monitors across various dimensions. I recommend setting up at least one monitor of type 'AWS Service' and another of type 'Linked Account' if your organization has multiple accounts. When you receive an alert, the Anomaly Details dashboard is a great starting point for root cause analysis.
Use Cost Explorer for root cause detection
Cost Explorer is a powerful tool for understanding and analyzing cloud spend. It’s recommended to schedule weekly and monthly reviews to analyze cost behavior from the previous period. A solid understanding of how to use Cost Explorer is essential for anyone managing Cloud FinOps. Knowing how to apply dimensions and filters to handle various root cause analysis scenarios and analyzing trends to derive actionable steps are key skills. Being well-versed in Cost Explorer can make the Cloud FinOps journey smoother for organizations of any scale.
Share your approach and suggestions for Cloud FinOps in the comments below. I'd love to try them out!
Top comments (0)