How Sedai Turned Cloud Infrastructure Into a Strategic Advantage

At most companies, the cloud is not built with a clear strategy from day one. Instead, the cloud becomes the product of countless decisions, made by different teams, at different times, to handle different tasks.

One team picks AWS for website hosting. Another spins up a GCP instance for data analytics. A new integration must run on Azure. After a few years, you’re running a complex multi-cloud, with 40 different services and thousands of resources.

It’s chaos. And no one has visibility over everything.

So at Sedai, we asked ourselves: How can we eliminate this chaos in the cloud? And what would it mean for the cloud to become a true strategic advantage?

This is an especially pressing problem, because so much of cloud spend is completely wasted. In 2026, 73% of companies run hybrid or multi-cloud, and cloud waste has ticked up to 40%. Getting out of that chaos requires a strategic philosophy for how companies build their clouds. 

At Sedai, we believe the modern cloud must be built to become a command center, where one team has full visibility across every provider and every service, optimization happens autonomously, and cost data informs engineering strategy.

Companies can no longer treat the cloud as infrastructure to simply monitor; instead it must become a strategic initiative that informs how applications are built, scaled, and optimized.

Here’s our philosophy for building your cloud as a strategic initiative: 

1. Your platform should run cohesively across any cloud and environment
2. Your visibility should tell you where to act, not just what you're spending
3. Optimization must be grounded in real application behavior & context
4. Your cloud should continuously evolve

That's the cloud Sedai was built for, and what we've built for ourselves, and we’ve seen it pay off.

Run Across Any Cloud, Any Environment

A strategic cloud must be built to run across any cloud or environment. We built our platform for multi-cloud observability from day one because we knew the market demanded it; a majority of our customers run on multiple clouds, covering Azure, GCP, and AWS.

That meant building Sedai to be fault tolerant, resilient, & designed to recover from failure.

To do this, we built Sedai on Kubernetes & an abstraction layer. We chose Kubernetes because it’s inherently cloud-agnostic by design and deployable anywhere. But Kubernetes alone wasn't enough. We needed to handle different services for different clouds, and Kubernetes doesn't abstract those differences. It just runs the containers. 

“The cloud is a living system that requires continuous adaptation, validates release impact, & makes safe decisions in real time.”

Hari Chandrasekhar

VP of Engineering

That meant if we hardcoded Sedai to depend on specific cloud services, it would break the moment it was deployed in a different environment. 

To solve this, we built an abstraction layer where individual components can be swapped per environment without touching the core application. Essentially, what runs on Kubernetes today can flip to a managed cloud service tomorrow, or be fully self-hosted for air-gap installations.

As organizations continue to race against growing costs, the flexibility of Sedai in conjunction with customers’ cloud readiness creates a single pane of observability, regardless of cloud provider.

By ingesting and normalizing billing data from each provider, organizations can get an accurate picture of how they’re spending across clouds. And this extends to Sedai’s autonomous optimization: with cross-cloud Kubernetes support built in, Sedai can manage EKS, GKE, & AKS workloads by container or clusters.

Spend Visibility That Tells You Exactly Where to Act

Our platform was built for multi-cloud cost visibility, and we believe high-level spend reports don’t provide the visibility needed to uncover & reduce hidden costs.

Instead, Sedai breaks down costs granularly: you can drill down into individual clusters, VMs, serverless functions, & services. This level of detail provides broader insight to what each environment costs, down to the namespace & component level. 

Engineers can easily benchmark these against each other, and know where exactly to put their engineering effort & toil to fix any inefficiencies or broader issues.

We experienced this when we were monitoring our own platform for inefficiencies, especially in how infrastructure was allocated across internal environments and services.

Without granular visibility, it was harder to see where capacity was actually needed, where resources were over-allocated, and which parts of the system could be scaled more precisely.

But with Sedai monitoring our own infrastructure, we got a clearer view of where resources were actually needed and where they could be scaled down or repurposed. By exposing more direct performance metrics and breaking broad services into more specific components, we could make more precise scaling decisions instead of scaling broad services indiscriminately.

This is how we strive to build our own strategic cloud: granular-level visibility pinpoints exactly where issues are occurring, all without the added toil of digging through dashboards and infrastructure code.

Optimize Around App Context & Behavior

The blueprint for the strategic cloud cannot stop at visibility. Knowing exactly where your money is going is only useful if you can act on it safely, and that means building in enterprise context and application understanding from the start.

Automated optimization tools rely on fixed rules that can fail when application behavior changes. For example, a tool might react to a CPU spike and recommend a change, but if the spike falls outside its predefined rules, that change could cause a production outage.

This forces engineering teams to spend their time manually validating changes instead of shipping. This is where Sedai’s context graph comes in.

From the start, Sedai builds a context graph of your cloud, including the topology of your environment, the dependencies between applications & cloud services, and how your infrastructure actually connects.

That context extends all the way down to Kubernetes, where Sedai maps deployments, pods, autoscalers, & their relationships to understand how workloads behave in practice.

By analyzing this context alongside your app’s historical load patterns, Sedai makes autonomous optimizations that are safe & effective.

You can’t have a strategic cloud where engineers spend their time validating changes. A strategic cloud is one where the platform understands your applications well enough to act on your behalf, confidently, safely, & without manual intervention.

That's the standard we hold ourselves to. And it's the standard we build toward for every customer.

Close the Loop With Release Intelligence

A key part of Sedai’s intelligence is understanding that app behavior changes with every new software release. An optimization platform that understands your application today, but fails to understand the impact of a deployment tomorrow, is incomplete. 

This is where Sedai’s Release Intelligence provides the crucial feedback loop that makes a strategic cloud possible.

While Sedai’s context graph builds a deep historical model for workloads, Release Intelligence analyzes the impact of change in real time. For every deployment, Sedai provides a per-deployment scorecard that analyzes the impact on cost, latency, & error rates. 

This gives teams immediate, data-driven feedback on whether a change improved or degraded the service. It connects development activity directly to operational outcomes.

This creates a powerful cycle:

1. Sedai understands the baseline behavior of a service, enabling safe, autonomous optimizations
2. When deploying a new version, Release Intelligence measures its precise impact on performance & cost.
3. New data continuously refines the application's behavior model, ensuring future optimizations are based on the most current reality

This is what elevates the platform from simply making a one-time recommendation to becoming a continuously learning system that adapts alongside apps. 

The strategic cloud continuously evolves. By measuring the impact of every change and release, it learns what works, adapts over time, and makes optimization a natural part of software delivery.

A Strategic Cloud Is an Autonomous Cloud

Simply put, you need an autonomous system to operate the cloud in 2026.

At Sedai, we believe the modern cloud should do more than react to problems or surface generic utilization data. It should drive strategy through portability, actionable visibility, and continuous optimization based on real-world behavior.

That’s how we built Sedai’s own infrastructure, and we’ve been able to drive real strategy & product decisions because of it.

At one point, Sedai identified that we were overspending on metric scraping for a customer’s AWS environment. The issue wasn’t just resource consumption, it was architectural. 

Our exporter component was stateful, meaning it handled scheduling, execution, & resiliency internally. That made the system expensive to run and difficult to scale precisely. The solution was to separate those responsibilities entirely.

Build Your Strategic Cloud

See how Sedai can turn your cloud from chaos into strategy.

Book a Demo

Now, a centralized state manager determines which scraping jobs should run and when, while stateless workers execute those jobs independently. By decoupling orchestration from execution, Sedai can calculate exactly how many workers are needed for a given workload and scale only what’s necessary.

The result was a more resilient & cost-efficient system. Failed workers can simply restart without carrying recovery logic inside the app itself, and scaling decisions became significantly more precise.

More importantly, it reinforced our core belief: When infrastructure continuously learns from real application behavior, optimization becomes a strategic advantage instead of a reactive task.

The Future of the Cloud Is Autonomous

As cloud environments become more complex, teams can no longer rely on static data & isolated recommendations to manage them effectively.

The cloud is no longer something teams configure once and maintain manually; it’s a living system that requires continuous adaptation, validates release impact, & makes safe decisions in real time. 

That’s why we believe the future of cloud management is autonomous.

This doesn’t mean removing engineers from the loop, but rather removing the repetitive operational burden that prevents them from focusing on higher-value work.

The truth is, you simply can't manage a modern cloud by asking different teams to optimize individual resources. The scale is too big. The complexity is too great.

If you’re ready to build your strategic cloud, schedule a demo to see Sedai in action.