Open for FinOps as Code Real Examples
The most complete edition regarding open source, tutorials and implementations
Victor Garcia
August 03, 2025
Together with CloudThrottle
How to Automate Cloud Cost
Tools like AWS Budgets, Azure Cost Management, or GCP's native tools offer only basic alerts and per-account visibility. You're forced to manage budgets in silos.
What’s missing?
Centralized control
Real-time budget feedback
Automated workflows
Cross-account enforcement
This leads to reactive decisions, late-stage alerts, and missed financial targets—especially in the public sector, education, or fixed-price contracts.
Born from federal contracting challenges, CloudThrottle is not just another dashboard. It's a proactive cloud budget management solution designed to bring predictability, visibility, and control to multi-account cloud environments.
AWS
Optimize Kubernetes on AWS: Graviton & Spot Guide with KEDA & Karpenter
The video teaches viewers how to cut Kubernetes expenses by up to 45% through two main strategies.
First, it shows how to use AWS Spot Instances, which are cheaper compute resources that AWS offers at reduced prices. Second, it covers AWS Graviton processors, which provide better performance for the money compared to traditional processors.
The tutorial demonstrates practical tools for managing these cost savings. Karpenter helps automatically scale your Kubernetes workloads up and down based on demand. KEDA works alongside it to make sure your applications respond properly to changing workload requirements.
AWS Principal Specialist Solutions Architect Christian Melendez leads the demonstration. He walks through real performance tests and shows step-by-step how to implement these money-saving techniques in actual Kubernetes environments.
The video runs for about 34 minutes and includes hands-on examples of migrating workloads to Graviton instances and setting up automatic scaling systems.
OPEN SOURCE
Plan your Costs in IaC with OpenInfraQuote
OpenInfraQuote is a free tool that helps you figure out how much your cloud setup will cost before you build it.
The tool works in two simple steps. First, it looks at your Terraform files and matches each resource to pricing information. Then it adds up all the costs and shows you a clear breakdown of what each piece will cost per month.
You can use it two ways. Check costs before making changes by running it on your Terraform plan files. Or see what your current setup costs by running it on your existing state files.
The output is easy to read and shows you exactly which resources cost what. It even tells you if costs went up or down compared to before, and highlights new or removed resources.
FINOPS EVENTS
Mastering AI Economics
AI workloads are powerful but they’re also expensive. In this webinar, we’ll break down practical strategies to manage, develop, and reduce unit costs across your AI stack.
Learn how top engineering and FinOps teams are aligning performance with budget by optimizing architecture, tracking true cost per model, and using practical insights to stay ahead of runaway spend.
What we’ll cover:
Understanding unit costs: What they are, why they matter, and how to track them
Cost-efficient architecture: Design patterns and trade-offs that lower compute and storage bills
Data & model strategy: How to optimize what you train, when, and where
FinOps for AI: Building transparency and accountability into fast-moving AI teams
Speakers:
Vaibhav Sharma
David Gross
Alon Savo
Host: Victor Garcia
September 9th - 6:00 PM CEST / 10AM EST
OPEN SOURCE
AWS Anomaly Guardian: Real-time Detection & Response
The system uses many AWS tools working together. API Gateway collects information. Kinesis handles lots of data flowing in. Lambda functions process everything quickly. Machine learning models from SageMaker spot the strange patterns.
The system can handle over 10,000 events every second. It spots problems in less than 100 milliseconds. It scales up or down based on how busy things get.
It watches for different types of weird behavior. Too many requests from one person. Money amounts that don't make sense. Someone logging in from two countries at the same time. Strange browsing patterns. System slowdowns.
Setting it up is pretty simple if you know AWS. You can use automated scripts or do it step by step. The system creates dashboards to show what's happening. It sends alerts when things go wrong.
The machine learning part uses several smart methods. Isolation Forest finds outliers. Statistical Process Control watches for changes. Deep learning spots patterns over time. It keeps learning from new data and feedback.
Carlos included detailed instructions for using the system. API endpoints for sending data and getting results. Ways to create and manage rules. Webhooks to connect with other systems.
This solution helps companies catch fraud, protect APIs from attacks, and monitor IoT devices. It reduces false alarms and catches real problems faster than humans could.
CLOUD PROVIDERS
AWS Released Document DB Serverless
That’s basically it, BUT! You have all the updates from all providers in case you missed any.
AWS
A Cheaper Way to Run Your AI Training Jobs! AWS Batch now directly supports Amazon SageMaker training jobs, eliminating the need for separate management.
Amazon DocumentDB (MongoDB-compatible) launches in serverless mode
Don't Need a Whole GPU? Now You Don't Have to Buy One! New G6f instances offer fractional GPUs
Microsoft Azure
Making Sense of Marketplace and FinOps! Microsoft's Office Hours session highlighted how the Microsoft Marketplace enhances FinOps practices
Google Cloud
CLOUD PROVIDERS
Navigating Cloud Commitments: What to Know Before You Dive
A cost accountant turned cloud expert shares hard-won lessons about buying cloud services the smart way.
The biggest choice you'll face with cloud providers is how to pay. You can go on-demand and pay for what you use, or make commitments and pay for what you think you'll use.
The trade-off is simple but scary. Commitments can save you 30% to 70% on costs, but you're locked in for 1 to 3 years. Break up early and you still pay the full bill.
Level 1 - Start Simple
Begin with hourly spend commitments on compute services. You don't need to know which region or machine type you'll use. Just commit to spending a certain amount per hour on virtual machines.
Level 2 - Get Specific
These older-style commitments lock you to specific regions and machine families. They're called Reserved Instances on AWS, Reservations on Azure, and Resource Based CUDs on Google. Only consider these for databases or if you use third-party management tools.
Level 3 - Capacity Planning
The most complex option involves committing to processing capacity rather than just machines. Services like DynamoDB need you to predict how many database reads and writes you'll need per hour. This requires deep partnership with your engineering teams.
Hidden Level - Short-term Analytics
Some analytics services let you make short-term commitments measured in hours or months instead of years. Good for teams with steady data processing needs.
The author's key advice rings throughout. Talk to your developers and engineers before making any commitment. They know what's coming down the pipeline that could make or break your savings plan.
📺️ VIDEO
Expert Masterclass: FOCUS Success Case in Azure
In this Masterclass, we teach you how to design an event-driven architecture based on Azure Functions that processes billing information and divides costs using a combination of general rules and specific allocations that can be managed and pre-set in advance. Using FOCUS and Power BI
🎖️ MENTION OF HONOUR
Optimizing Cost and Scaling with Karpenter at Freshworks
Freshworks had a problem with their Kubernetes setup. Even though Kubernetes was supposed to help them use their computer resources better, they were wasting a lot of power. Only 35% of their processors and 54% of their memory were being used.
They were using Cluster Autoscaler to manage their servers, but it wasn't working well. They barely used any spot instances, which are cheaper servers that Amazon offers. This meant they were spending too much money. So they decided to try Karpenter instead.
Testing Karpenter
Freshworks tested Karpenter in four main areas. They wanted to make sure it was fast, reliable, easy to monitor, and cheap to run.
The tests showed that Karpenter could replace interrupted servers in less than 2 seconds. It handled clusters with 10,000 applications and 200 servers without problems. It also gave them better information about what was happening in their system.
Setting it up right
They created template files that work for most of their needs. These templates help different teams use Karpenter without making mistakes. The templates also prevent teams from choosing expensive server types by accident.
The results
After switching to Karpenter, Freshworks saved 15-25% on their server bills. They also increased their use of cheaper spot instances from 2% to 8%. In one case, they reduced the number of servers needed from 24 to a much smaller number while still running all their applications.
When not to use Karpenter
Karpenter works best for applications that change size often. For databases and other services that stay the same size all the time, the benefits are smaller. These steady workloads don't need the smart sizing that Karpenter provides.
Professional Spotlight
Carlos Filho
FinOps as Code Crack
A recent discovery, Carlos is publishing great content where he dives into technical implementation (FinOps as Code) articles. Love his style, and his contribution to the community.
That’s all for this week. See you next Sunday!
Join The Largest FinOps Online Event of 2025
Join FinOps professionals at the FinOps Weekly Summit 2025 and discover how to:
Transform from reactive fire-fighting to strategic leadership — Learn the proven frameworks that top practitioners use to turn cloud cost management into a competitive advantage
Master AI-powered cost optimization — Get exclusive access to the latest automation tools and techniques that can reduce your cloud spend by up to 40% while accelerating innovation
Build bulletproof unit economics — Walk away with actionable playbooks for calculating true cloud ROI and proving business value to executives who control your budget
"But I don't have time for another webinar..."
This isn't another generic webinar. This is the only major FinOps event left in 2025, featuring battle-tested strategies from companies managing billions in cloud spend.
October 23rd, 2025 | 4:00 PM - 8:00 PM CEST
Limited seats available