Fleets
Fleets automatically add or remove nodes in your Cloudfleet Kubernetes Engine (CFKE) cluster as your workload demands change. CFKE lets you deploy, run, and monitor applications in your own cloud provider account using a unified Kubernetes cluster. You can take advantage of existing discounts and credits while avoiding vendor lock-in. CFKE automatically selects optimal instance types and orchestrates workloads within your account - or bursts to other providers when needed.
Fleets introduction
Fleet defines in what cloud provider the infrastructure will be provisioned. A single Fleet contains configuration information for one or more cloud provider accounts where Cloudfleet can provision nodes for your workloads.
Every cluster in CFKE can have one or more Fleets and simultaneously run workloads on infrastructure from multiple Fleets. This allows you to take advantage of the best pricing and performance for your workloads, as well as to avoid vendor lock-in.
CFKE manages infrastructure provisioning based on the workloads deployed in the Kubernetes cluster. It uses scheduling preferences defined in the workload specifications to determine where resources should run and how much infrastructure is needed. CFKE then provisions the required nodes using the cloud provider’s APIs. It also handles the full node lifecycle - including scaling, upgrades, and deletions - so users can focus on their applications instead of managing infrastructure. To learn more about node-autoprovisioning, please refer to the Node Autoprovisioning documentation.
Supported providers
Node autoprovisioning with Fleets is available for select Cloud Service Providers (CSPs) integrated with Cloudfleet. These providers are selected based on reliability, performance, service breadth, and - most importantly - customer demand. Currently, Cloudfleet supports the following CSPs:
- Amazon Web Services (AWS): AWS offers industry-leading 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): GCP provides advanced infrastructure including 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: Hetzner delivers high-performance virtual machines powered by AMD EPYC and Intel Xeon processors, fast NVMe storage, scalable private networking, and a cost-effective pricing model - ideal for developers and enterprises seeking reliable European cloud infrastructure.
You can learn how to configure Fleets here.
We are continuously working to expand support for additional providers, aiming to deliver broader service options and geographical coverage to meet our customers’ evolving needs. If there’s a specific provider you’d like to see supported, please let us know. We value your feedback and prioritize platform growth based on customer demand.
Alternatively, you can use self-managed nodes to connect any cloud provider to CFKE. For details, refer to the Self-Managed Nodes documentation.
Supported accelerators
Cloudfleet offers a wide range of instance types equipped with AI/ML accelerators. These accelerators, primarily specialized hardware such as GPUs, are specifically designed to boost the performance of machine learning workloads. Our platform features over 250 different instance types, each tailored to suit the diverse requirements of AI and ML projects.
In our API, we have streamlined the process of selecting the appropriate accelerator for your needs. By standardizing the names of accelerator manufacturers and models, we have made it easier for users to identify and choose the right infrastructure instance type. This approach reduces confusion and accelerates the selection process, allowing users to focus more on their projects and less on navigating infrastructure complexities.
Below is the list of currently supported accelerators along with their available memory configurations:
Model | Providers | Memory configurations (GiB) |
---|---|---|
AMD V520 | aws | 8 |
AWS Inferentia | aws | 8 |
AWS Inferentia2 | aws | 32 |
Intel Gaudi HL-205 | aws | 32 |
NVIDIA 141GB | gcp | 141 |
NVIDIA A10 | aws | 24 |
NVIDIA A100 | aws, gcp | 4, 8 |
NVIDIA H100 | aws, gcp | 8 |
NVIDIA H200 | aws | 141 |
NVIDIA K80 | aws | 12 |
NVIDIA L4 | aws, gcp | 22, 24 |
NVIDIA L40S | aws | 44 |
NVIDIA T4 | aws | 16 |
NVIDIA V100 | aws | 16, 32 |
Qualcomm AI100 | aws | 15 |
Xilinx VU47P | aws | 8 |
Xilinx VU9p | aws | 64 |
This list is regularly updated to reflect new accelerators as they are integrated into our platform. For the latest information and detailed specifications of each accelerator, please refer back to this page in the future.
Node regions →