Fargate
Definition
Fargate is an AWS serverless container platform that allows users to run containers without managing servers or clusters, streamlining application
Use Cases
- Amazon: Running internal microservices and batch-style container workloads without managing EC2 worker nodes — Teams deploy containers using Amazon ECS with the Fargate launch type, defining task CPU/memory and networking, and letting AWS provision and scale the underlying compute (Reduced operational overhead from patching and managing worker nodes, faster provisioning for new services, and pay-for-use compute aligned to workload demand)
- Expedia Group: Operating containerized microservices with less infrastructure management overhead — Uses AWS container services (commonly ECS) and can run tasks on AWS Fargate to avoid managing EC2 capacity for certain services and jobs (Improved agility for teams by shifting responsibility for server management to AWS and enabling more consistent, repeatable deployments of container workloads)
Provider Equivalents
- AWS: AWS Fargate
- Azure: Azure Container Apps
- GCP: Google Cloud Run
- OCI: OCI Container Instances
Frequently Asked Questions
- What's the difference between AWS Fargate and Amazon ECS?
- Amazon ECS is the container orchestrator that schedules and manages your containers. AWS Fargate is a compute option for running ECS tasks (or EKS pods) without managing the underlying servers. In practice, you often use ECS for orchestration and choose Fargate as the launch type so you don’t manage EC2 instances.
- When should I use AWS Fargate?
- Use Fargate when you want to run containers without managing EC2 instances or Kubernetes worker nodes—especially for microservices, APIs, event-driven workers, scheduled jobs, and spiky or unpredictable traffic. It’s a good fit when operational simplicity matters more than deep control over the host OS, custom AMIs, or specialized instance types.
- How much does AWS Fargate cost?
- Fargate pricing is based on the vCPU and memory you allocate to each task or pod and how long it runs, plus any additional charges for related services (for example, load balancers, data transfer, and logging). Costs typically increase with higher CPU/memory reservations, longer runtimes, and always-on services. You can reduce cost by right-sizing task resources, using autoscaling, and stopping tasks when not needed.
Category: containers
Difficulty: intermediate
Related Terms
See Also