Manual processes, hidden costs and a culture of caution

Overprovisioning as a defense mechanism

For a platform serving global football traffic, caution is understandable. A major live match, breaking transfer news, or World Cup moment can create sudden and unpredictable traffic spikes. In that environment, the cost of underprovisioning is not only technical. It can directly affect the fan experience.

Historically, to protect performance during high-traffic events, OneFootball’s engineers often allocated more CPU and memory than services typically required. The logic was sound: it was safer to provide extra capacity than to risk service degradation during a critical moment in a football match.

But the cumulative effect of that caution was significant. Resources sat idle between peaks, infrastructure costs increased, and different teams lacked a consistent way to understand whether each service was configured appropriately for the right balance of efficiency, resilience, and production confidence.

Over time, overprovisioning became a defensive operating model. It helped reduce immediate risk, but it also made it harder to identify hidden inefficiencies, missed optimization opportunities, and configuration patterns where reliability could be preserved with a better cost profile.

Manual optimization that could not keep pace

Kubernetes resource management at OneFootball was largely manual. Engineers reviewed dashboards, analyzed service behavior, and adjusted CPU and memory allocations by hand. This work required time, context, and repeated judgment calls from both platform engineers and service owners.

The process was valuable, but it could not scale with the speed and volatility of live-sports demand. Traffic during a football event can spike and subside within minutes. Demand can also change based on global tournaments, national competitions, country-specific events, breaking news, and match outcomes that are impossible to predict precisely in advance.

A manual review cycle could not continuously account for those changes, especially across a growing Kubernetes environment serving global audiences. It also made it difficult to apply meaningful timeframes consistently, because the relevant traffic window for a service could vary significantly depending on the event, market, or competition.

Manual optimization also created inconsistency. Different teams could make different decisions based on different levels of Kubernetes expertise, risk tolerance, or available time. As a result, resource configuration was not always driven by a consistent, data-backed understanding of actual service behavior.

From observability to actionable cost and resiliency insights

OneFootball already had observability in place. Teams could monitor services, investigate behavior, and understand production signals through their existing tooling. The challenge was not a lack of observability, but the gap between observing what was happening and continuously turning those signals into actionable Kubernetes optimization decisions.

Cloud provider dashboards gave OneFootball visibility into overall infrastructure spend, but they did not provide the level of service-level insight the platform team needed to balance cost, efficiency, and resilience. Teams could see that infrastructure costs existed, but it was harder to understand which services were driving those costs, whether resources were being used efficiently, and where optimization could happen without increasing operational risk.

This lack of optimization intelligence made decision-making difficult. Without a clear view of requests, limits, usage patterns, cost impact, and resiliency signals at the service level, identifying the right tradeoff often required manual investigation and interpretation.

That is where resiliency insights and alerting became the “cherry on the cake.” They did not replace observability. They complemented it by helping the platform team identify what might otherwise be missed or overlooked: services that were overprovisioned, workloads with inefficient resource profiles, or configuration patterns where resilience and cost could be better balanced.

Together, these challenges described an engineering organization operating carefully, but defensively. As OneFootball continued to grow and prepare for major global football events, the platform team needed a more automated, transparent, and reliable way to optimize Kubernetes without asking engineers to choose blindly between cost efficiency and production confidence.

The solution: automated intelligence, financial clarity and expert support

When OneFootball started leveraging PerfectScale by DoiT, the objective was not simply to introduce another dashboard. The platform team already had observability in place. What they needed was a way to continuously translate Kubernetes behavior into actionable optimization decisions across cost, efficiency, autoscaling, and resilience.

PerfectScale provided that missing layer of intelligence. It helped OneFootball move from manual investigation and defensive overprovisioning toward automated, data-backed optimization that could support the volatility of global football traffic.

Visibility that changed the conversation

From day one, PerfectScale gave the platform team a more precise service-level view of how Kubernetes resources were requested, consumed, and translated into cost. Engineers could understand not only CPU and memory utilization, but also the financial and operational consequences of resource allocation decisions.

That visibility changed the way infrastructure conversations happened across the organization. Instead of discussing Kubernetes optimization only as a technical concern, the platform team could connect resource decisions to business impact, production confidence, and cost efficiency.

This was particularly important in a live-sports environment, where services must be prepared for sudden spikes while still avoiding unnecessary waste between peaks. PerfectScale made those tradeoffs easier to understand by showing where resources were underused, where configuration could be improved, and where optimization could happen without compromising reliability.

Continuous automated rightsizing

PerfectScale replaced manual resource review cycles with continuous, automated rightsizing across OneFootball’s Kubernetes environment. Instead of relying on engineers to periodically review dashboards and adjust CPU and memory allocations by hand, PerfectScale provided ongoing recommendations and automation based on actual service behavior.

This was critical for traffic patterns driven by football. A major match, transfer story, push notification, or World Cup moment can change demand quickly and unevenly across services. Manual optimization could not reliably keep pace with those shifts, especially when the relevant traffic window might vary depending on the competition, country, or event.

With PerfectScale, OneFootball could optimize workloads more continuously and consistently. Services could be tuned based on real usage patterns rather than assumptions, helping the platform team reduce waste while preserving the confidence required for production workloads.

The platform also helped surface resiliency insights that complemented OneFootball’s existing observability stack. These insights gave engineers an additional layer of visibility into configuration patterns, resource profiles, and optimization opportunities that could otherwise be missed or overlooked.

Hands-on support and continuous collaboration

Beyond the platform itself, DoiT’s broader support ecosystem played an important role in building trust and accelerating adoption. OneFootball had access to customer success, post-sales support, solution engineering, and Kubernetes expertise, giving the platform team a clear path for both strategic guidance and day-to-day questions.

This support was especially valuable because OneFootball was not looking for a generic optimization model. The platform team needed PerfectScale to fit the way OneFootball operates Kubernetes, manages production risk, and approaches platform engineering.

“I spent a lot of time digging into PerfectScale from a platform engineering perspective. The value was not only seeing optimization recommendations, but understanding why they made sense, how they mapped to real workload behavior, and how they helped us balance cost, autoscaling, and resiliency tradeoffs safely. The regular syncs with the PerfectScale team also helped us validate edge cases and shape the integration around how OneFootball actually operates Kubernetes.” Andrea Benfatto, Platform/Cloud Runtime Engineering Manager, OneFootball

Regular PerfectScale syncs helped both teams iterate on the integration over time. These sessions gave OneFootball a way to provide feedback, validate recommendations, discuss edge cases, and shape the setup around its own operational needs. As a result, the integration became more aligned with OneFootball’s workflows, traffic patterns, and platform expectations.

DoiT’s solution engineers and Kubernetes specialists worked closely with OneFootball’s platform team to ensure recommendations were technically sound, operationally safe, and aligned with production requirements. This hands-on collaboration helped reduce risk, accelerate decision-making, and reinforce confidence in the system.

For OneFootball, the value came from the combination of automation, service-level visibility, resiliency insights, and expert support. PerfectScale did not replace existing engineering judgment or observability. It complemented them with continuous optimization intelligence, helping the platform team make better Kubernetes decisions at scale.

OneFootball is continuing to expand its use of PerfectScale by DoiT, moving beyond automated optimization into deeper governance, control, and event-readiness across its Kubernetes environment.

A key next step is the integration of PerfectScale recommendations and policies into OneFootball’s infrastructure-as-code workflows. With the recent PerfectScale CRDs integration, the platform team can bring optimization closer to the GitOps approach it already aims to follow: defining, reviewing, and managing Kubernetes configuration through version-controlled workflows.

This is an important evolution for OneFootball. As the company prepares for major global football events, including the 2026 FIFA World Cup, the platform team needs optimization practices that are not only automated, but also auditable, repeatable, and aligned with how engineering teams already operate.

By managing optimization policies through Kubernetes-native resources, OneFootball can apply more granular control over resource allocation, autoscaling behavior, and resiliency tuning. This creates a stronger connection between service ownership, production requirements, and infrastructure efficiency.

The goal is not only to reduce cost. It is to make Kubernetes optimization a governed part of the platform operating model: one that supports global traffic, helps teams prepare for unpredictable demand, and gives engineers confidence that efficiency improvements are being made safely.

Ongoing PerfectScale syncs with DoiT will continue to play an important role in this evolution. These sessions give OneFootball a way to validate new capabilities, provide feedback from real production usage, and shape the integration around the company’s specific platform needs. Over time, this collaboration helps ensure that PerfectScale continues to fit the way OneFootball operates Kubernetes at scale.

As PerfectScale becomes part of DoiT’s broader platform, OneFootball also sees opportunities to further strengthen its infrastructure strategy with additional capabilities, deeper optimization intelligence, and expert support.

For OneFootball, PerfectScale is no longer just a tool for Kubernetes rightsizing. It is becoming part of how the platform team designs, governs, and prepares its infrastructure for the next generation of global football moments.