Successful Kubernetes Case Studies. Part 2
Harshit Mehndiratta
Harshit Mehndiratta
August 23, 2020
8 minutes read

Successful Kubernetes Case Studies. Part 2

We have already discussed various real-world examples of organizations across industries that are migrating to containers and Kubernetes in Part 1 of Successful Kubernetes Case Studies.

In this part, we will provide you seven more real-world examples of organizations that are leveraging containers and Kubernetes as a powerful combination for pushing workloads and data to cloud environments.


The largest Financial network in the world, Bloomberg, turned to Kubernetes in 2017. Bloomberg, on an everyday basis, used to process more than hundreds of billions of data. There are more than 14,000 different applications on the terminal alone, which were up and running. Over the years, the Bloomberg infrastructure was dependent on spinning up a lot of VM’s for scaling and processing. But that did not provide app teams enough flexibility for their application development and scaling needs.

After evaluating many Docker offerings and different firms such as Cloud Foundry, Mesosphere, they have chosen Kubernetes to align with their business needs. Bloomberg aimed to make better use of their existing hardware to deploy new applications and services to users while freeing up the burden of operational tasks on developers.

As a result, after migrating to Kubernetes, Bloomberg reaped many benefits from the platform. Developers became more productive; fewer errors were there in deployment, improved service request handling, and seamless automation. They were also able to efficiently use the hardware to the point where they can get close to 90 to 95% utilization rates.


Airbnb migrated to Kubernetes when they need to support over 1,000 engineers who were concurrently deploying essential services for the company. Airbnb needed to scale continuous delivery horizontally. The aim was to make that continuous delivery available to the company’s 1,000 or so engineers so they could add and configure new services as fast as possible.

The net result came out to be quite astonishing, Kubernetes enabled Airbnb to add a layer of abstraction to their configurations from higher-level primitives using namespaces and automated validations. This was important for the automation of common workflows for engineers and teams using the same tools across all environments.

Kube-gen, a template rendering tool for Kubernetes, helped Airbnb take service parameters (defined in a single YAML file) and generate the complete Kubernetes services manifests containing all the necessary boilerplate configurations.

Furthermore, a newly created service’s git repository includes applications, infrastructure, and CI/CD pipeline configurations - auto-filled with sensible defaults. Service configurations include a new service skeleton repository that provides configuration validation at the build time.

As of now, with Kubernetes Airbnb can do over 500 deployments per day on average while supporting concurrent deployment and configuration of over 250 critical services.


Vodafone Group is one of the world’s leading global telecommunications company that provides technology services. They have expertise in a variety of Internet of Things(IoT) and connectivity products for both consumers and businesses, as well as mobile financial services and digital transformation in emerging markets.

Vodafone struggled with an old, monolithic architecture that had incurred high levels of complexity, interdependency, and a substantial upgrade deficit. As a result, they could no longer maintain it and needed to start fresh with new technology.

In 2016, Vodafone launched a new digital strategy to deliver the best customer services. The digital strategy includes implementing cloud-native software and container orchestration platform Kubernetes, to best enable local markets to share assets across different parts in the world.

Vodafone has chosen Giant Swarm, which provided them a highly automated control plane to manage Kubernetes and the stack. The product came with features that increased Vodafone’s peace of mind while adopting moving to the cloud. There were no vendor lock-in issues, which provided an easy pivot in case of failure. It also eliminated the challenges of managing a Kubernetes environment that was not well-known to the Vodafone staff.

Right now, a blueprint for cloud-native projects partnered with Giant Swarm is created. This blueprint is now being adopted throughout Vodafone Group, especially with Kubernetes projects that are moving from the development phase to production.

Expedia Group

Expedia Group is an American online travel services company for consumers and small businesses. Its websites, which are travel search engines, primarily include,, trivago, and Vrbo.

Expedia is derived from a combination of “exploration” and “speed”, which is the reason the company began using container architecture via Amazon EC2 Container Service (ECS) in 2015 to deploy and run micro-services in a public cloud environment efficiently.

Expedia also made use of Kubernetes as their container orchestration platform for their unit. It currently runs hundreds of its applications on Kubernetes in the AWS cloud, providing services such as hotel bookings for customers.

After the launch of Elastic Kubernetes Service (EKS) by Amazon, all major teams of Expedia Group, including data science, customer experience, and analytics, all began to leverage container-based workloads.

The primary aim for deploying the container-based architecture and Kubernetes was its greater portability across multiple environments and accelerated speed for delivering new products and services. While containers are spinning up in seconds, the amount of time which is needed to ship a new feature to customers has significantly reduced using container architecture and Kubernetes.

Thousands of applications are running in production on EKS right now across multiple AWS cloud regions, enabling Expedia to provide enhanced services for travelers and partners across the globe.


Travel Tech Giant Amadeus provides global distribution systems that enable online travel services companies like Expedia and Kayak to book flights, rooms, and car bookings. Amadeus has been working predominantly with cloud computing platforms such as Google Cloud. They have shifted one of their most critical application, ‘Master Pricer’, to the Google Cloud Platform.

The next critical step was to migrate to Kubernetes. The overall aim was to use Kubernetes with Google Cloud to provide a broader business perspective for shifting on-premises deployments to the public cloud and cut down on infrastructure costs to meet seasonal scaling demands.

In late 2016 the organization started its first airline availability app migration to Kubernetes in production with Red Hat’s OpenShift platform. The original plan was to move another application related to hotel booking first. Still, the airline availability app was preferred to be moved to the public cloud first to serve the growing demands of clients better.

With Kubernetes, Amadeus was able to handle several thousand transactions per second, which are deployed in multiple data centers across the world. “It’s a whole new workload that we couldn’t have done otherwise” says Eric Mountain, senior expert at distributed systems at Amadeus IT group.


Tinder’s move to Kubernetes was to drive Tinder Engineering towards container architecture and immutable deployment so engineers could focus more on their code than operation tasks such as infrastructure, application build, and deployment.

Tinder’s engineering team faced a struggle of scale and stability with AWS ECS. They often used to wait several minutes for EC2 instances to come online during scaling, which was only seconds with Kubernetes.

Tinder’s Journey to Kubernetes Migration started in early 2018. There were various stages of the migration process. Starting from containerizing their services and deploying them to a Kubernetes-hosted environment. Then in Late 2018, they started moving their legacy services to Kubernetes. By March of 2019, the Tinder Platform was completely running on Kubernetes consisting of 200 services, 1,000 nodes, 15,000 pods, and 48,000 running containers.

During the migration, Tinder developed a robust in-house infrastructure team with great familiarity with designing, deploying, and operating large Kubernetes clusters. Tinder’s entire engineering organization now has sufficient experience to containerize and deploy applications on Kubernetes.

After Kubernetes, containers at Tinder also served traffic within seconds as opposed to minutes. Scheduling multiple containers on EC2 instance also provided horizontal density, which promoted cost savings in 2019 compared to the previous year.


Launched in 2008, Spotify is one of the largest music streaming subscription service that has grown to over 200 million monthly active users across the world. Spotify aims to empower creators and enable an immersive listening experience for their customers. They have containerized their microservices and managed it through an in-house container orchestration service called Helios in 2014.

By late 2017, Spotify decided to migrate from Helios to Kubernetes, which was backed by a huge community of thousands of developers. Kubernetes comes with a huge ecosystem behind it that gave a lot of confidence in transition. Spotify was aiming to benefit from autoscaling, better resource utilization, self-healing, faster operations, and development.

The migration started in parallel with Helios running. Spotify wants to run all stateless services on Kubernetes for which they have created migration phases. In the first phase, they decided to go small, experimenting with only running one service on a Kubernetes cluster, moving up to three services on a shared cluster for a few days. This will ease out the transition and would only affect a few teams.

With the first phase complete, Spotify moved on to the alpha phase, asking teams to volunteer services for running on Kubernetes. During this phase, the services ran on Helios and Kubernetes, and Spotify can roll back to Helios if an incident occurred. Then comes the beta phase of self-service migration, in which teams that want to migrate to Kubernetes can follow the documentation and fully move to Kubernetes.

During the general availability phase, any new service at Spotify will be deployed only on Kubernetes. Spotify employed capabilities such as one-click migration, autoscaling, metrics, and audits.

Spotify benefited hugely with Kubernetes. According to their teams, now they have less of a need to focus on manual capacity provisioning, and more time to focus on delivering features. Spotify running on Kubernetes can take about 10 million requests per second and benefits greatly from autoscaling, says James Wen. Site Reliability Engineer. Teams can now create a new service in a few minutes. Kubernetes has also improved CPU utilization compared to compute instances on Helios.

Kubernetes-Driven Future

One of the more fascinating parts of these different Kubernetes success stories is their consistency irrespective of their industry. Organizations of all sizes are wrestling with a change in the manner development is done in large and small businesses. Most of the above-defined Kubernetes organizations want to adopt Kubernetes to accelerate their software development cycles, cut costs, and provide more customer satisfaction to their clients. It’s safe to say that Kubernetes has reached a level where any organization that neglects to find a good pace with containers and Kubernetes will battle to thrive in a cloud-driven world.