Run containers without managing servers, with per-second billing, multi-container groups on shared hosts, and optional GPU support for ML inference workloads
FluffyStack toolsJurisdictional exposure
Attributes
- SLA Uptime
- 99.9%
- GPU Support
- Yes
- Max Memory
- 16 GB
Sub-services (2)
Container Groups
Multi-container pods on the same host
GPU Container Instances
GPU-enabled containers for ML workloads
Compliance & Certifications
This service is attested for the following frameworks. Always verify with the provider before relying on a specific compliance posture.
Where this runs
Sovereign regions (13)
- Australia Central · CanberraAzure Australia Government
- Australia Central 2 · CanberraAzure Australia Government
- US Gov Virginia · VirginiaAzure Government (US)
- US Gov Arizona · ArizonaAzure Government (US)
- US Gov Texas · TexasAzure Government (US)
- US DoD East · VirginiaAzure Government Secret (US)
- US DoD Central · IowaAzure Government Secret (US)
- China North (Beijing) · BeijingMicrosoft Azure China (21Vianet)
- China East (Shanghai) · ShanghaiMicrosoft Azure China (21Vianet)
- China North 2 · BeijingMicrosoft Azure China (21Vianet)
- China East 2 · ShanghaiMicrosoft Azure China (21Vianet)
- China North 3 · HebeiMicrosoft Azure China (21Vianet)
- China East 3 · ShanghaiMicrosoft Azure China (21Vianet)
Commercial regions (60)
Europe (21)
- Austria East
- Belgium Central
- Denmark East
- Finland Central
- France South
- France Central
- Germany North
- Germany West Central
- Greece Central
- North Europe
- Italy North
- West Europe
- Norway East
- Norway West
- Poland Central
- Spain Central
- Sweden Central
- Switzerland West
- Switzerland North
- UK West
- UK South
North America (13)
- Canada East
- Canada Central
- Mexico Central
- West US
- East US 3
- North Central US
- Central US
- West US 3
- South Central US
- East US
- East US 2
- West US 2
- West Central US
South America (3)
- Brazil Southeast
- Brazil South
- Chile Central
Asia (13)
- East Asia
- South India
- Jio India West
- West India
- Jio India Central
- Central India
- Indonesia Central
- Japan West
- Japan East
- Malaysia West
- Southeast Asia
- Korea South
- Korea Central
Oceania (3)
- Australia East
- Australia Southeast
- New Zealand North
Middle East (5)
- Israel Central
- Qatar Central
- Saudi Arabia Central
- UAE Central
- UAE North
Africa (2)
- South Africa West
- South Africa North
Tags
Equivalent services on other platforms
Highly secure and reliable container orchestration service with EC2 and Fargate launch types, service discovery via Cloud Map, and deep integration with ALB, IAM, and CloudWatch
Serverless compute engine for containers that runs tasks on ECS or EKS with per-second billing, no cluster management, and task-level CPU and memory sizing
Fully managed service to deploy containerised web apps and APIs from source code or container images without managing clusters, load balancers, or scaling rules — automatic HTTPS, auto-scaling between zero and hundreds of concurrent requests, and VPC connectivity
Fully managed serverless container platform built on Kubernetes, KEDA, and Dapr for running microservices and APIs without managing cluster infrastructure, with scale-to-zero, traffic splitting, revisions, and built-in service discovery
Serverless execution of containerised tasks to completion on Cloud Run, with parallelism, task arrays, retries, timeouts up to 24 hours, and the same revision / VPC / IAM controls as Cloud Run services — purpose-built for nightly batch, data pipelines, and scheduled work
Fully managed serverless container platform that runs stateless HTTP services and jobs from container images with automatic scaling to zero and per-request billing
Serverless container runtime built on Kata Containers with native Kubernetes API, second-level billing, auto-scaling to zero, and seamless bursting from CCE-managed nodes for spill-over workloads
Container service that runs containers as first-class OpenStack resources via Docker, Kata, or runC backends — the platform equivalent of ECS / Container Instances for OpenStack-based private clouds, separate from the Magnum project (which provisions Kubernetes clusters)
Serverless container runtime that runs OCI Container Image-compliant workloads without managing VMs or Kubernetes, with per-second billing, up to 64 OCPUs per instance, instance pools for horizontal scale, and native VCN networking