When Oleksandr Leonhard reflects on the project that marked a turning point in his cloud infrastructure career, he doesn’t just talk about servers or architectures. He talks about momentum — for the product, for the clients, and for the business itself.
At the time, Oleksandr was part of the engineering team at DashDevs, where they were building an internal platform for business process management tailored to IT companies.
Don’t want to miss the best from TechLatest ? Set us as a preferred source in Google Search and make sure you never miss our latest.
The idea was simple: give organisations a way to automate workflows, visualise operations, and improve internal coordination. But as the platform matured and client interest grew, it became clear that the existing infrastructure wasn’t ready to handle the pressure — financially or technically.
“We were starting to hit walls,” Oleksandr says. “The infrastructure was expensive to maintain, particularly as the user base grew and the data load increased. Our compute costs were unpredictable, and we spent a lot of engineering time just maintaining the system’s stability under load.”
Recognising that this would only become a bigger problem as the company scaled, Oleksandr proposed a full assessment of their cloud strategy. The goal was to re-architect the system with cost-efficiency, scalability, and speed of development in mind.
This eliminated much of the manual work they had previously handled, saving time and reducing the risk of human error. “With RDS, we could rely on AWS to handle the essential maintenance,” Oleksandr explains. “It immediately reduced our operational burden.”
Next, the team turned to their compute strategy. Instead of keeping a static number of virtual machines running regardless of demand, they implemented Auto Scaling for EC2 instances.
This allowed the system to automatically adjust the number of running servers based on real-time traffic, ensuring that resources were only used when needed. It helped smooth out the peaks and valleys of user activity without incurring unnecessary costs during off-peak periods.
However, the most significant transformation came when the team began to integrate serverless computing via AWS Lambda and AWS Fargate. By moving specific workloads to Lambda functions, they were able to eliminate the cost of idle servers entirely.
This shift from a traditional server-based model to a function-as-a-service approach meant that they only paid for actual execution time, not for servers sitting idle waiting for requests.
“Lambda fundamentally changed how we thought about scaling,” Oleksandr says. “We no longer had to estimate the number of servers we’d need ahead of time. Instead, we built services that could scale automatically and cost-effectively.”
But adapting the application to serverless required more than a technical shift — it demanded a new approach to development.
To speed up the process and avoid reengineering for each client adaptation, the DashDevs team created their own internal framework on top of AWS services.
This framework allowed them to quickly develop and deploy new features and workflows, each one built as a modular Lambda function.
“With the framework, we could roll out new functionality in weeks instead of months,” Oleksandr notes. “That was a huge advantage, especially when clients asked for customizations or new capabilities. We could respond faster than our competitors, and it also cut down development costs significantly.”
After proving the platform internally, DashDevs began offering it to other companies, adapting the solution to fit each client’s unique business processes.
Thanks to the framework’s modular design, every deployment was quicker and more cost-efficient than the last. This adaptability turned the initial infrastructure upgrade into a new revenue stream for DashDevs.
During the re-architecture process, the team also evaluated whether to use Terraform, particularly given its flexibility for infrastructure-as-code across cloud providers.
While powerful, Terraform’s broader scope came with complexity that wasn’t necessary for their AWS-focused stack.
Instead, they opted for AWS CloudFormation, which provided tighter integration with AWS services and simplified onboarding for new engineers.
“CloudFormation gave us everything we needed without the additional overhead that comes with Terraform,” says Oleksandr. “It streamlined our deployments and reduced the learning curve for the team. For our goals, the trade-off made sense.”
Of course, shifting to serverless and auto-scaling wasn’t without challenges. The team encountered concurrency limits in AWS Lambda, especially during testing phases when sudden bursts of requests could push the system beyond its configured thresholds.
To keep financial control tight, they implemented AWS Cost Explorer along with a detailed tagging system to monitor costs across different services, environments, and client projects.
This visibility was essential not only for budgeting but also for making continuous improvements to their infrastructure.
In the end, the results were clear. DashDevs reduced infrastructure costs by 25%, delivered a faster, more reliable product, and significantly decreased the time required to customize solutions for new clients.
Perhaps more importantly, the team reclaimed the bandwidth to focus on product innovation rather than infrastructure firefighting.
“We didn’t just cut costs,” Oleksandr reflects. “We created a system that empowered us to scale both technically and commercially.
The combination of AWS services and our internal framework allowed us to move faster, spend less, and stay ahead of the curve.”
Today, that work continues to pay dividends. The platform Oleksandr helped re-engineer has become a foundation for multiple client solutions, each tailored yet cost-effective thanks to the groundwork the DashDevs team laid.
“When you’re building infrastructure, you’re not just solving today’s problems,” he says. “You’re setting up your company’s ability to grow without being trapped by operational costs. That’s what AWS allowed us to do, and it’s an approach I carry with me in every project I take on.”
This story was originally published on 3 September 2022.
Enjoyed this article?
If TechLatest has helped you, consider supporting us with a one-time tip on Ko-fi. Every contribution keeps our work free and independent.
