FinOps in action: How we helped Reiseklar slash Azure costs and maximize performance
At NoA, we make sure our clients’ digital tools work smarter, not harder. That includes keeping cloud expenses in check—because they can snowball quickly if you’re not paying attention. That’s exactly what happened to Reiseklar, a Norwegian travel app. Their Azure costs were climbing, but performance still needed to stay top-notch. So, we made a bold move: cut the bloat while maintaining speed. The result? A 65.9% reduction in costs—without breaking a thing.
About the Project
Reiseklar is a travel app developed by the Norwegian Ministry of Foreign Affairs. It keeps travelers in the loop with real-time notifications and guidelines. For example, if a major incident happens where you’re staying, the Norwegian Ministry of Foreign Affairs will ping you through the app, SMS, or email. Simply put, think of it as a pocket-sized travel advisor with a power of a search engine.
The challenge
Reiseklar put performance first (totally valid), but they weren’t looking to burn cash on cloud costs either. However, the bigger the project, the higher the cloud costs—unless you stay on top of them. Reiseklar were facing:
- Too many cloud resources
- The risk of performance taking a hit if they scaled down
- No clear auto-scaling setup to balance cost vs. Performance
- An API setup that was costing way more than it should
As Reiseklar’s trusted tech partner, we wanted to make sure their infrastructure runs smoothly, efficiently, and, most importantly, cost-effectively. We aimed to find the sweet spot between top performance and smart spending.
Our Approach
First Up: Trimming the Fat in the Test Environment
Test environments are important but don’t need to be running at full power 24/7. Reiseklar’s was overprovisioned—big time. The quick fix? Cut instances from 10 to 3 and boom—32% cost savings. But wait, there’s a catch…
A deeper dive showed that switching tiers would nix deployment slots—a big no-go for staged releases. So, we stayed put and found savings elsewhere instead.
Fine-Tuning Application Insights
Cutting monitoring costs sounded tempting, but we weren’t about to lose critical performance insights. Instead, we fine-tuned the setup, adding even more detailed traces to the test environment. As a result, we saw a tiny increase in monitoring costs but way better performance tracking and issue detection. Totally worth it.
Optimizing the Production Environment App Service Plan
Let’s talk about one of the biggest budget-eaters in cloud hosting: the App Service Plan. This bad boy can account for up to 40% of total infrastructure costs—so, naturally, we had to take a closer look.
Reiseklar was running on an older, legacy plan (yikes), and we spotted a golden opportunity to save some cash. By switching to a newer plan, we scored:
- A slight reduction in Azure Compute Units (ACU) per vCPU (210 to 195)
- A slight increase in memory allocation (3.5GB to 4GB)
Win-win! Performance stayed solid, but costs dropped by 23%. And just to be extra sure, we analyzed CPU utilization—it was hanging below 5%, so reducing the tier was a no-brainer.
Implementing Smart Auto-Scaling Policies
Now, we get it—performance and stability were Reiseklar’s top priorities. They weren’t looking to slash costs at all costs (pun intended). But hey, efficiency? Always a good idea.
Our deep dive into usage metrics revealed a key turning point: we could safely reduce the number of instances without risking slowdowns—if we had smart auto-scaling in place for high-traffic moments.
So, we set up realistic scaling policies to:
Optimal performance during traffic spikes
Reduced costs during periods of lower demand
Azure API Management (APIM) Optimization
Usually, API Management accounts for 10–30% of cloud costs. But for Reiseklar it was a jaw-dropping 60%. Yep, that needed fixing.
We rolled up our sleeves and got to work, analyzing every possible way to trim the fat without hurting API performance. Our approach included:
- Comparing feature sets across different pricing tiers
- Identifying interchangeable functionalities across tiers
- Evaluating cache size requests to optimize API performance
- Simulating various load scenarios to measure performance impact
After crunching the numbers, we downgraded Reiseklar’s APIM tier to the Basic plan—massively cutting costs while keeping API efficiency intact.
Load Testing for Performance Assurance
Optimizing cloud costs is great, but not if it wrecks performance. So, we put everything through serious load testing—simulating peak traffic, stress-testing API endpoints, and making sure Reiseklar’s infrastructure could handle the heat.
The verdict? Stable, high performance even after cost optimizations. No slowdowns, no hiccups—just a leaner, meaner cloud setup that’s built to last.
The Results
Key results:
- 65.9% cost reduction in Azure hosting fees
- Zero impact on service response times
- More efficient resource allocation without sacrificing key features
Key optimizations implemented
- Adjusted service plans to lower, more cost-effective tiers
- Implemented auto-scaling policies for optimal resource usage
- Reduced DTU allocations for SQL databases
- Optimized API Management usage to avoid unnecessary costs
- Strengthened monitoring to prevent performance issues
In Their Own Words
If you want a deep dive into your system, I wholeheartedly recommend this service. It’s advertised as FinOps, but to do that, they need to do an in-depth test on the complete backbone. It delivers an overview of your services—more than just whether you’re in the right tier or have enough memory allocations. In effect, you get complete due diligence on your entire backbone solution. You save money, but more importantly, you can confidently report that your service is running effectively. Peace of mind —priceless.
Christian Birkely
Product owner, Reiseklar
Norwegian Ministry of Foreign Affairs
Conclusions
At NoA Ignite, cloud cost optimization isn’t a one-and-done deal. It’s an ongoing process. By continuously tweaking configurations, fine-tuning auto-scaling, and staying sharp on resource usage, businesses can cut costs without losing speed.
For Reiseklar benefited from our smart approach meant by instant savings and long-term efficiency. Now, they can focus on innovation, growth, and making travel safer—without sweating their cloud bill.
