How Wildflower cut Kubernetes costs by 50%
with advanced automation
“We’ve already reduced the node count in our cluster by half. I was looking for a 30% reduction. It’s going to be more like 65%.”
Blake Mitchell
VP of Engineering at Wildflower
“It used to take several hours out of my week. Sometimes entire days. Now it’s completely off my plate. But it’s not just saving time. It’s giving me capabilities I didn’t have before.”
Blake Mitchell
VP of Engineering at Wildflower
“Before, Kubernetes would overload nodes with more pods than they could handle. Zesty solved that problem by continuously adjusting the appropriate resource requirements on our workloads.”
Blake Mitchell
VP of Engineering at Wildflower
Customer
Brief
Wildflower is a digital health company dedicated to connected maternity care. Its mobile app bridges the gap between patients, providers, and payers, offering counseling and remote monitoring to detect pregnancy risks early, improve care quality, and lower healthcare costs. The company’s engineering operations are built entirely on AWS, running Kubernetes-based infrastructure with roughly 800 pods.
Key
Challenges
Wildflower’s Kubernetes costs were growing faster than their business. Manual monitoring consumed valuable time and still led to overprovisioning, inefficiencies and silent workload failures.
Key
Results
Zesty’s Pod Rightsizing enabled Wildflower to instantly cut Kubernetes costs by up to 50%, eliminate infrastructure failures, and automate resource management with zero effort.
The Challenge:
For years, Wildflower struggled with one of Kubernetes’ most fundamental challenges: figuring out how much CPU and memory each workload actually needed. It was a guessing game, leading to inefficiencies and instability.
The company experiences steady usage with daily and weekly traffic patterns, peaking in the mornings and during business hours, with a noticeable dip around Christmas. But with no dynamic and historical visibility into Kubernetes resource usage, Wildflower managed infrastructure reactively, adding nodes to the cluster after encountering issues like throttling or out-of-memory errors, which occurred several times a month.
Manual tuning took hours each week and still left them exposed to inefficiencies. They struggled with jobs silently failing to start, due to inaccurate scheduling and overloaded nodes. Blake, VP of Engineering at Wildflower, explained, “Our periodic data jobs would silently stop running for hours or even days, and we wouldn’t realize it until much later.”
In addition, Wildflower’s AWS costs were increasing faster than business growth, prompting a push to identify spending inefficiencies, particularly in Kubernetes, where they had no clear understanding of the value of their spend.
Zesty’s Solution:
Wildflower initially considered tools like Cloudwatch, but it required a steep learning curve and significant time investment, which Wildflower couldn’t afford, so they turned to Zesty.
The company was already seeing success with Zesty’s Commitment Manager, which they described as the best software they’ve used: effortless to implement and instantly effective in lowering their AWS bill through Reserved Instances and Savings Plans.
When they discovered Zesty’s Pod Rightsizing, it immediately stood out as the first solution that could finally automate this process. For the first time, they could stop guessing, gain visibility into both real-time and historical usage, and let the system set accurate resource requirements. That clarity was exactly what they had been missing, and it’s what convinced them to give Zesty a try.
The onboarding process was simple and hands-on, with engineers helping set up the solution live over a call. Once deployed, it gave the team real visibility into how resource needs changed over time, and it eliminated the need for manual monitoring.
The Result:
Zesty Pod Rightsizing helped Wildflower automatically adjust pod resource requirements, and as a result, they were able to cut their cluster node count by 50%, far beyond their original 30% savings target. With automation fully in place, they expect savings to reach 65%.
Blake explained that with Zesty Pod Rightsizing, he finally gained visibility into how resource needs evolve over time, saying the best part is that he’s no longer “blind” and no longer has to constantly tweak configurations by hand.
The solution did more than just save time by offloading the routine task of resource tuning. It also unlocked new capabilities he never had before. Many optimization tasks were so time-consuming that they simply didn’t get done. But with automation in place, he gained the ability to optimize resources far more efficiently, while freeing time to upgrade and secure their Kubernetes foundation.
With Zesty, Wildflower has gone from reactive infrastructure scaling to a proactive, automated model, preventing them before they happen. Workload stability improved, thanks to accurate CPU and memory requests that enabled Kubernetes to schedule correctly and prevent cascading failures.
Finance leadership quickly recognized Zesty’s solution as a major contributor to reducing AWS spend, standing out as one of the clear wins in cost discussions, much like the Commitment Manager had before it, and they were genuinely excited about the results.
Blake shared that with Zesty, he had found the missing piece and filled the long-standing gap in Kubernetes with the visibility and automation that had been missing for years.
Customer
Brief
Wildflower is a digital health company dedicated to connected maternity care. Its mobile app bridges the gap between patients, providers, and payers, offering counseling and remote monitoring to detect pregnancy risks early, improve care quality, and lower healthcare costs. The company’s engineering operations are built entirely on AWS, running Kubernetes-based infrastructure with roughly 800 pods.
Key
Challenges
Wildflower’s Kubernetes costs were growing faster than their business. Manual monitoring consumed valuable time and still led to overprovisioning, inefficiencies and silent workload failures.
Key
Results
Zesty’s Pod Rightsizing enabled Wildflower to instantly cut Kubernetes costs by up to 50%, eliminate infrastructure failures, and automate resource management with zero effort.
“We’ve already reduced the node count in our cluster by half. I was looking for a 30% reduction. It’s going to be more like 65%.”
Blake Mitchell
VP of Engineering at Wildflower
For years, Wildflower struggled with one of Kubernetes’ most fundamental challenges: figuring out how much CPU and memory each workload actually needed. It was a guessing game, leading to inefficiencies and instability.
The company experiences steady usage with daily and weekly traffic patterns, peaking in the mornings and during business hours, with a noticeable dip around Christmas. But with no dynamic and historical visibility into Kubernetes resource usage, Wildflower managed infrastructure reactively, adding nodes to the cluster after encountering issues like throttling or out-of-memory errors, which occurred several times a month.
Manual tuning took hours each week and still left them exposed to inefficiencies. They struggled with jobs silently failing to start, due to inaccurate scheduling and overloaded nodes. Blake, VP of Engineering at Wildflower, explained, “Our periodic data jobs would silently stop running for hours or even days, and we wouldn’t realize it until much later.”
In addition, Wildflower’s AWS costs were increasing faster than business growth, prompting a push to identify spending inefficiencies, particularly in Kubernetes, where they had no clear understanding of the value of their spend.
“It used to take several hours out of my week. Sometimes entire days. Now it’s completely off my plate. But it’s not just saving time. It’s giving me capabilities I didn’t have before.”
Blake Mitchell
VP of Engineering at Wildflower
Wildflower initially considered tools like Cloudwatch, but it required a steep learning curve and significant time investment, which Wildflower couldn’t afford, so they turned to Zesty.
The company was already seeing success with Zesty’s Commitment Manager, which they described as the best software they’ve used: effortless to implement and instantly effective in lowering their AWS bill through Reserved Instances and Savings Plans.
When they discovered Zesty’s Pod Rightsizing, it immediately stood out as the first solution that could finally automate this process. For the first time, they could stop guessing, gain visibility into both real-time and historical usage, and let the system set accurate resource requirements. That clarity was exactly what they had been missing, and it’s what convinced them to give Zesty a try.
The onboarding process was simple and hands-on, with engineers helping set up the solution live over a call. Once deployed, it gave the team real visibility into how resource needs changed over time, and it eliminated the need for manual monitoring.
Zesty Pod Rightsizing helped Wildflower automatically adjust pod resource requirements, and as a result, they were able to cut their cluster node count by 50%, far beyond their original 30% savings target. With automation fully in place, they expect savings to reach 65%.
Blake explained that with Zesty Pod Rightsizing, he finally gained visibility into how resource needs evolve over time, saying the best part is that he’s no longer “blind” and no longer has to constantly tweak configurations by hand.
The solution did more than just save time by offloading the routine task of resource tuning. It also unlocked new capabilities he never had before. Many optimization tasks were so time-consuming that they simply didn’t get done. But with automation in place, he gained the ability to optimize resources far more efficiently, while freeing time to upgrade and secure their Kubernetes foundation.
With Zesty, Wildflower has gone from reactive infrastructure scaling to a proactive, automated model, preventing them before they happen. Workload stability improved, thanks to accurate CPU and memory requests that enabled Kubernetes to schedule correctly and prevent cascading failures.
Finance leadership quickly recognized Zesty’s solution as a major contributor to reducing AWS spend, standing out as one of the clear wins in cost discussions, much like the Commitment Manager had before it, and they were genuinely excited about the results.
Blake shared that with Zesty, he had found the missing piece and filled the long-standing gap in Kubernetes with the visibility and automation that had been missing for years.
“Before, Kubernetes would overload nodes with more pods than they could handle. Zesty solved that problem by continuously adjusting the appropriate resource requirements on our workloads.”
Blake Mitchell
VP of Engineering at Wildflower
