Multi-cloud Kubernetes clusters

Run workloads across one or multiple cloud providers within a single Kubernetes cluster. Cloudfleet unlocks new cost optimization strategies and streamlined application deployments without the need to manage complex automation tooling.

Get started Schedule demo

CFKE - Cloudfleet Kubernetes Engine

Future of the compute layer of the internet

Multi-cloud, or simply consuming cloud services from multiple providers, is becoming a de facto standard. It can be a conscious decision by a company to make use of multiple vendors to distribute their workloads geographically, optimize costs, or access best-of-breed technologies. In larger organizations, it’s also common to become multi-cloud over time as a result of mergers and acquisitions or due to individual departments making independent purchasing decisions.

Where you run your workloads should be based on your business needs, not the limitations of your infrastructure vendor. If done right, multi-cloud delivers significant benefits without introducing additional operational overhead.

Cloudfleet Kubernetes Engine (CFKE) is a fully managed Kubernetes service that allows you to run your applications on any cloud provider, in any region, from a single control plane. Cloudfleet runs the Kubernetes control plane in a managed and secure environment, ensuring that all critical components are always available and up to date. No matter the size of the cluster or whether your infrastructure is in an on-premises data center in Asia or an AWS region in the US, you can focus on your application’s configuration and only pay for the resources the application uses.

/images/lp/multi-cloud/cloudfleet-multi-cloud-kubernetes.png

What’s best for you and your customers

/images/lp/multi-cloud/multicloud-multi-cloud-providers.png

At Cloudfleet, we believe that the future of the compute layer of the internet is not bound by vendor limitations: Intel CPUs and Nvidia GPUs are largely the same. Nowadays, raw server power is undifferentiated, and we’ve built Cloudfleet Kubernetes Engine to let you run workloads virtually anywhere from a single control plane. With Cloudfleet, you can pick the cheapest spot instance from hyperscalers, complement your primary cloud provider with cheaper niche players, or even use bare metal.

Cloudfleet offers access to the largest global infrastructure catalog compared to any single cloud provider, and it continues to expand rapidly. This gives you the flexibility to deploy your applications and workloads in the cloud infrastructure closest to your target users, ensuring optimal performance. Whether you’re running high-throughput applications or those with strict low-latency requirements, Cloudfleet allows you to choose the cloud environment that best meets your needs, providing superior support for a broad range of applications.

Best-of-breed cloud services

/images/lp/multi-cloud/multicloud-multi-cloud-services.png

Cloudfleet provides a seamless experience with the same level of integration as native Kubernetes services offered directly by cloud providers. You can access the cluster from your CI/CD pipelines, use autoscaling, storage interfaces, or other managed services without compromising security.

Services running on Cloudfleet Kubernetes Engine have native access to any service on any cloud (e.g., BigQuery on GCP, Active Directory on Azure, S3 on AWS) without hardcoding credentials, using CFKE’s Kubernetes Service Accounts. This enables you to easily mix and match services from multiple clouds by unifying networking, identity, and authorization policies across all environments.

Uniform identity and access management

Cloudfleet maintains a single permissions management model for both cluster administrators and end users. This means you can use your internal identity provider (IdP) as a single source of truth for authentication and a unified Kubernetes role-based model for fine-grained resource access control. There is no need to synchronize or copy permissions between vendors and different systems.

Available on all plans, including the free tier, Cloudfleet provides SSO and integration with your IdP via SAML and LDAP protocols, supporting products like Microsoft Active Directory, Google Directory, or Okta.

The combination of Kubernetes-native JWT tokens and Cloudfleet OIDC endpoints allows you to reduce security exposure and establish access to external services without managing secrets. This enables greater standardization and governance across vendors while simplifying day-to-day business functions.

Taking care of Day-2 operations

Cloudfleet covers the most challenging aspects of managing the Kubernetes control plane, including networking, security, updates, scaling, monitoring, and high availability. By simplifying platform management, Cloudfleet enables organizations to easily overcome Day-2 operational barriers and get up and running in minutes instead of weeks or months.

Cloudfleet Kubernetes Engine comes with a state-of-the-art stack of foundational components pre-installed, pre-configured, and tested. We take care of networking, securing your traffic with site-to-site VPNs, and operating cloud-specific controller managers and persistent volume adapters. The best infrastructure is the one you don’t have to actively manage - Cloudfleet provides the easiest way to run global Kubernetes clusters.

apiVersion: v1
kind: Pod
metadata:
  name: example-app
spec:
  nodeSelector:
    cfke.io/provider: gcp
    topology.kubernetes.io/region: europe-west4
  containers:
  - name: example-app
    image: nginx
    resources:
      requests:
        memory: "256Mi"
        cpu: "1000m"
      limits:
        memory: "512Mi"
        cpu: "2000m"

Kubernetes manifest example

Specific cloud provider and region

This manifest deploys an Nginx pod on a Kubernetes cluster, ensuring it runs only on nodes in the ‘europe-west4’ region of Google Cloud (GCP) with at least 256 MiB of memory and 1 CPU core.

apiVersion: apps/v1
kind: Deployment
metadata:
  name: example-app
spec:
  selector:
    matchLabels:
      app: nginx
  replicas: 2
  template:
    metadata:
      labels:
        app: nginx
    spec:
      nodeSelector:
        karpenter.sh/capacity-type: spot
      affinity:
        nodeAffinity:
          requiredDuringSchedulingIgnoredDuringExecution:
            nodeSelectorTerms:
              - matchExpressions:
                - key: "cfke.io/provider"
                  operator: "In"
                  values: ["gcp", "aws"]
        podAntiAffinity:
          requiredDuringSchedulingIgnoredDuringExecution:
            - labelSelector:
                matchExpressions:
                  - key: app
                    operator: In
                    values:
                      - nginx
              topologyKey: cfke.io/provider
      containers:
        - name: nginx
          image: nginx:1.24
          resources:
            requests:
              cpu: 50m
              memory: 56Mi

Kubernetes manifest example

Spread spot across AWS and GCP

This manifest deploys an Nginx application with two replicas, ensuring they run on spot instances across both AWS and GCP. It uses node affinity to restrict scheduling to nodes labeled as either GCP or AWS, while pod anti-affinity ensures that replicas are spread across different cloud providers to improve resilience.

apiVersion: apps/v1
kind: Deployment
metadata:
  name: cuda-vector-add
spec:
  selector:
    matchLabels:
      app: cuda-vector-add
  template:
    metadata:
      labels:
        app: cuda-vector-add
    spec:
      nodeSelector:
        cfke.io/accelerator-name: H100
      containers:
        - name: cuda-vector-add
          image: "nvidia/samples:vectoradd-cuda11.6.0"
          resources:
            limits:
              nvidia.com/gpu: "1"

Kubernetes manifest example

Cheapest Nvidia H100 across clouds

This manifest deploys a CUDA-based GPU workload on the cheapest available Nvidia H100 accelerator across multiple cloud providers. It ensures that the pod is scheduled only on nodes equipped with an H100 GPU using a nodeSelector and requests one GPU for execution. The container runs the NVIDIA Vector Add sample, demonstrating CUDA computation.

Cloudfleet solves the biggest challenges

Operating Kubernetes, especially at scale and in distributed environments, requires specialized skills and can be complex. Cloudfleet provides out-of-the-box features and capabilities to accelerate time to production for containerized workloads - on-premises, in the cloud, in hybrid deployments, or at the edge.

Global load balancing and ingress

Cloudfleet dynamically provisions and deprovisions load balancers based on cluster nodes. A single service can be served from multiple clouds and regions, ensuring high availability and resilience.

Data transfer out (DTO) cost optimization

All traffic in Cloudfleet clusters is routed using topology-aware load balancing, minimizing unnecessary latency and reducing the cost of distributed deployments.

Shared storage and persistent volumes

Cloudfleet clusters come with multi-site CSI, allowing you to provision persistent volumes of various types across multiple locations. You can also bring your own CSI driver.

Control plane latency and availability

Maintaining consensus and state management is critical for cluster stability. Cloudfleet ensures sub-millisecond latency between the Kubernetes API and the state database for optimal performance and reliability.

Cross-border networking and encryption

Nodes in Cloudfleet clusters connect via a secure, encrypted overlay network spanning multiple clouds and regions. This WireGuard-based network has minimal overhead, operates at high speed, and leverages state-of-the-art cryptography.

Single global Kubernetes control plane

Cloudfleet delivers truly global clusters - eliminating the need for a ‘cluster of clusters,’ separate clusters for each availability zone or data center, or complex ‘single pane of glass’ abstractions and vendor-specific solutions.

Cloudfleet infrastructure providers

Largest network of cloud providers

Cloudfleet offers access to the largest global infrastructure catalog of any single cloud provider, and it continues to expand rapidly. This gives you the flexibility to deploy applications and workloads in the cloud infrastructure closest to your target users, ensuring optimal performance. CFKE automatically provisions Kubernetes nodes across 12+ cloud providers and 1,300+ node types. You do not need to plan and provision capacity upfront - CFKE dynamically allocates compute capacity based on the workloads you deploy on Kubernetes.

Amazon Web Services (AWS)

Self-managed nodes  |  Node auto-provisioning

AWS offers cutting-edge infrastructure with Nitro-based EC2 instances for enhanced security and performance, scalable Elastic Block Store (EBS) volumes up to 64 TiB, and ultra-low-latency networking for HPC workloads.

Google Cloud Platform (GCP)

Self-managed nodes  |  Node auto-provisioning

GCP features advanced infrastructure with custom Tensor Processing Units (TPUs) for AI acceleration, high-performance Persistent Disks supporting up to 120,000 IOPS, and global load balancing capable of handling millions of requests per second.

Hetzner Cloud

Self-managed nodes  |  Node auto-provisioning

Hetzner Cloud provides high-performance virtual machines powered by AMD EPYC and Intel Xeon processors, fast NVMe storage, scalable private networking, and a cost-efficient pricing model - ideal for developers and enterprises seeking reliable European cloud infrastructure.

Oracle Cloud Infrastructure (OCI)

Self-managed nodes

OCI offers ultra-fast NVMe-based block storage with sub-millisecond latency, supports up to 1 million IOPS per instance, and provides a high-bandwidth RDMA network with 100 Gbps throughput for HPC and database workloads.

Microsoft Azure

Self-managed nodes

Azure delivers ultra-low-latency virtual machines powered by AMD EPYC and Intel Xeon processors, high-speed InfiniBand networking up to 400 Gbps for HPC, and durable Azure Ultra Disk Storage supporting up to 160,000 IOPS and 2,000 MB/s throughput per disk.

IBM Cloud

Self-managed nodes

IBM Cloud provides enterprise-grade infrastructure with support for single-tenant bare metal servers deployed in under 2 hours, up to 12 Tbps DDoS protection via global load balancing, and quantum computing access through IBM Quantum Systems.

Paperspace

Self-managed nodes

Paperspace, a cloud platform specialized in AI and machine learning, offers high-performance NVIDIA GPUs such as A100 and H100 for scalable training workloads, dedicated networking with up to 50 Gbps bandwidth, and fast NVMe storage for low-latency data access.

Cloudfleet Kubernetes adoption

Start your journey the way you want

Create a free Kubernetes cluster

Cloudfleet offers an always-free Basic plan for Kubernetes cluster with up to 48 CPUs.
Sign up now

Join the Cloudfleet Pilot Program

Get hands-on support from Kubernetes experts and access to exclusive resources.
Apply now

Request a Cloudfleet platform demo

We’re happy to show you around! Schedule a call with the Cloudfleet team to see CFKE in action.
Request demo

Multi-cloud Kubernetes FAQ

Multi-cloud refers to the strategy of using cloud services from multiple providers, such as AWS, Google Cloud, and Microsoft Azure, rather than relying on a single provider. This approach enables businesses to optimize costs, improve resilience, and avoid vendor lock-in by distributing workloads across different cloud environments.

Kubernetes is an ideal platform for multi-cloud deployments because it provides a consistent orchestration layer across different cloud providers. It abstracts infrastructure differences, allowing applications to be deployed, scaled, and managed uniformly regardless of where the underlying nodes are hosted. With features like declarative deployments, automated scaling, and workload portability, Kubernetes enables seamless multi-cloud operations while optimizing for cost, performance, and reliability.

Companies choose a multi-cloud strategy for several reasons, including cost optimization, resilience, flexibility, and compliance. Here are the key benefits:

  1. Avoiding Vendor Lock-In. Relying on a single cloud provider can lead to high switching costs and limited flexibility. Multi-cloud allows businesses to distribute workloads across providers and avoid dependency on one vendor’s pricing, features, or roadmap.
  2. Cost Optimization. Different cloud providers offer varying pricing models and discounts. By adopting a multi-cloud strategy, companies can leverage the most cost-effective services from different providers, reducing overall cloud expenses.
  3. Improved Resilience & High Availability. Outages happen - even with major cloud providers. Multi-cloud helps increase uptime and disaster recovery capabilities by distributing workloads across different cloud infrastructures. If one provider experiences downtime, services can failover to another.
  4. Best-of-Breed Services. Each cloud provider has its unique strengths (e.g., AWS for compute, Google Cloud for AI/ML, Azure for enterprise integrations). A multi-cloud strategy enables businesses to combine the best services from different platforms to build a more powerful and efficient system.
  5. Compliance & Data Sovereignty. Regulatory requirements often mandate that certain data must be stored within specific regions or jurisdictions. Multi-cloud allows companies to deploy workloads in different regions to comply with legal and security regulations.
  6. Performance Optimization & Latency Reduction. By deploying applications in multiple cloud regions closer to users, companies can reduce latency and improve performance for global customers.
  7. Innovation & Competitive Advantage. Using multiple clouds fosters innovation by enabling teams to experiment with new cloud-native tools and frameworks without being restricted to a single provider’s ecosystem.

While multi-cloud offers flexibility, resilience, and cost advantages, it also introduces several challenges:

  • Complex Management & Operations - Each cloud provider has its own tools and APIs, making centralized management difficult. Solutions like Kubernetes, Terraform, and cloud-agnostic management platforms help streamline operations.
  • Security & Compliance Risks - Varying security models across clouds increase the risk of misconfigurations and data breaches. Implementing zero-trust security, unified identity management, and standardized policies is essential.
  • Networking & Latency Issues - Ensuring seamless connectivity between clouds can be complex and expensive. Private interconnects (e.g., AWS Direct Connect, Google Cloud Interconnect) and service mesh solutions help optimize networking.
  • Data Management & Portability - Different storage formats and high data transfer costs make multi-cloud data management challenging. Using cloud-agnostic storage layers, replication strategies, and cost-efficient data movement mitigates this.
  • Skill Gaps & Expertise Shortages - Managing multiple clouds requires specialized knowledge across different platforms. Investing in cross-cloud training and automation reduces operational complexity.
  • Cost Management & Optimization - Hidden fees, data transfer costs, and resource inefficiencies can drive up expenses. Using FinOps best practices, cloud cost monitoring tools, and automated cost controls helps optimize spending.

Despite these challenges, Cloudfleet simplifies multi-cloud management with automation and governance, making it a powerful strategy for businesses looking to enhance reliability and scalability.

Cloudfleet Kubernetes Engine (CFKE) simplifies multi-cloud Kubernetes by automating the provisioning, scaling, and management of both the control plane and worker nodes across multiple cloud providers, regions, and availability zones.

  • High Availability & Fault Tolerance - CFKE ensures a resilient control plane by automatically detecting and replacing unhealthy nodes, applying patches, and distributing workloads across clouds to prevent downtime.
  • Fully Managed Kubernetes - CFKE removes the burden of managing servers by handling the API servers, persistence layers, and compute resources for you.
  • Serverless Compute for Containers - No need to manually manage nodes; CFKE provisions and scales worker nodes dynamically based on application demand.
  • Cost Efficiency - You only pay for the resources your applications use, eliminating unnecessary infrastructure overhead.
  • Security by Design - CFKE enhances security through built-in application isolation, reducing the attack surface and enforcing best practices.

By eliminating infrastructure complexity, CFKE allows you to focus on building and scaling applications without worrying about multi-cloud orchestration, security, or maintenance.

Cloudfleet enables you to create, deploy, operate, monitor, upgrade, and retire one or thousands of clusters with ease, across multiple regions and cloud providers. We continuously benchmark our Pro tier clusters to ensure they can handle large-scale, high-intensity workloads across hundreds or even thousands of nodes, maintaining performance and reliability at any scale.

Absolutely. Cloudfleet provides significant benefits even when used within a single cloud provider by enabling access to spot and preemptible instances, which offer up to 90% cost savings compared to on-demand pricing. Unlike traditional solutions, Cloudfleet optimizes infrastructure usage across all regions of a single CSP, ensuring you always get the best pricing and availability.

Additionally, Cloudfleet automates workload failover and interruption handling, keeping your applications operational even when spot instances are reclaimed. While Cloudfleet excels in multi-cloud deployments, using it within a single provider still helps reduce costs, improve resilience, and avoid vendor lock-in, making it a future-proof choice for scaling your infrastructure.

Managing data transfer costs in a multi-cloud environment requires a strategic approach to minimize unnecessary expenses. One effective way is to leverage cloud providers that offer free ingress and egress, ensuring that data movement between services doesn’t incur high fees. Additionally, using vendor-independent, scalable global storage can help keep data accessible across multiple clouds without frequent inter-cloud transfers, reducing both latency and costs.

It’s also essential to evaluate the cost-benefit of moving data before initiating transfers. Not all data needs to be continuously synchronized across clouds, so optimizing workflows to limit unnecessary movement can lead to significant savings. Having the right tooling in place, such as Cloudfleet, helps track, analyze, and optimize data transfer costs, ensuring efficient use of multi-cloud resources while keeping expenses under control.

Yes, Cloudfleet fully supports hybrid cloud deployments, allowing you to extend your Kubernetes clusters to on-premises infrastructure. By connecting your data center to Cloudfleet, you can seamlessly integrate existing workloads with cloud-native applications, enabling use cases like hybrid cloud bursting, disaster recovery, and edge computing.

Cloudfleet Kubernetes Engine (CFKE) is specifically designed for on-premises infrastructure, supporting hyperconverged, bare metal, and virtualized environments with a fully managed control plane. This ensures high availability, security, and ease of use, regardless of where your clusters are deployed - whether in a data center in Asia or a cloud region in the US.

Learn more about multi-cloud Kubernetes and Cloudfleet

Documentation

Use cases

Tutorials

Enterprise-ready container platform

SLAs and dedicated support

Data processing and service level agreements (SLAs), dedicated account manager

Identity federation

Flexible Single Sign-On (SSO), SAML support, and role-based access control (RBAC)

Enterprise-grade security

Governance, audit logging, and SOC 2 and ISO 27001 (pending) certifications

Contact sales

Built to empower. Designed to scale.

Multi-cloud

Comprehensive support for major cloud providers and regions, ready for multi-cloud deployments.

Unlimited scalability

Expand servers, increase storage, or migrate to a different provider with zero downtime - at the push of a button.

99.99% uptime

High availability is included in the open-source feature set, ensuring reliability for mission-critical applications.

Integrations

Monitor and manage your cloud infrastructure effortlessly with your preferred tools.

World-class support

Expert 24/7 support, available 365 days a year. We're here whenever you need us.

Built-in security

End-to-end encryption, isolated networking, dedicated VMs, and full compliance certifications.

Latest posts from Cloudfleet blog.

/images/post_hero/hero1.png
Cloud Native Tips & Tricks
Proxmox vs Kubernetes in 2025: understanding the differences and similarities
/images/post_hero/hero2.png
Cloudfleet Partner Network News
A Look into Google Axion Processors: Google’s New ARM-Based CPUs
/images/post_hero/hero3.png
Cloud Native Tips & Tricks
Kubernetes 1.31 Release: Key Features, Enhancements, and Deprecations
/images/photos/kubernetes-engineer.jpg

Get started with a free Kubernetes cluster

Create your free Cloudfleet Kubernetes cluster in minutes - no setup hassle, no cost. Get started instantly with the always-free Basic plan.

Get started →