AWS Fargate

Serverless compute engine for running containers without provisioning or managing servers. Lets teams deploy containerized workloads and focus on applications while the platform handles infrastructure lifecycle, capacity, scaling, and OS updates. Works with both ECS and EKS as the compute backend for orchestrated containers.

Used by enterprises such as BILL for infrastructure modernization at scale. Differentiates from self-managed nodes by removing server management, capacity planning, and cluster operations. Each task or pod runs in a dedicated, single-tenant runtime with workload isolation. Compliance certifications cover HIPAA, PCI, and FedRAMP.

Key features:

• Serverless compute for OCI-compliant containers with no servers to manage
• Compatible with ECS and EKS orchestration
• Per-task isolation with dedicated virtualization boundary per workload
• Integrations with CloudWatch Container Insights and third-party observability tools
• Stateful workloads via EFS integration for externalized storage
• Scale up to 16 vCPU and 120 GB memory per task
• Windows containers support for .NET migration without refactoring

Use cases:

Teams building web applications, APIs, and microservices deploy containerized workloads without operating cluster infrastructure. Data processing and batch jobs benefit from serverless parallel processing and integration with managed batch services. ML training and inference workloads scale compute on demand while avoiding overprovisioning. Migration scenarios include running Windows containers for legacy .NET applications without rearchitecting.