A Spot Fleet is a collection of EC2 instances (Spot and optional On-Demand) that launches and maintains capacity from multiple Spot pools, offering cost savings, automatic capacity maintenance, and flexible allocation strategies.
A Spot Fleet is a resource that manages a collection of Spot Instances (and optionally On-Demand Instances) that launch from multiple Spot Instance pools. Each pool is defined by an instance type, operating system, and Availability Zone. The Spot Fleet automatically requests capacity from the most cost-effective pools based on your allocation strategy and can maintain target capacity even when individual instances are interrupted [citation:2].
submitted– The fleet request is being evaluated and EC2 is preparing to launch instances [citation:2]active– The fleet has been validated and EC2 is attempting to maintain target capacity [citation:2]modifying– The fleet request is being modified (only formaintaintype fleets) [citation:2]cancelled_running– Fleet is cancelled but instances continue running [citation:2]cancelled_terminating– Fleet is cancelled and instances are terminating [citation:2]cancelled– Fleet is cancelled with no running instances [citation:2]
lowestPrice – Launches instances from the pool with the lowest price. For scale down, terminates instances with the highest price per unit [citation:3][citation:5]
diversified – Distributes instances across all pools, reducing risk of simultaneous interruption [citation:3][citation:5]
capacityOptimized – Launches from pools with optimal available capacity for the number of instances being launched [citation:3][citation:5]
capacityOptimizedPrioritized – Uses priority (if set) on a best-effort basis while optimizing for capacity first [citation:3]
Cost Optimization: Automatically selects cheapest Spot pools while maintaining target capacity
Fault Tolerance: Diversifies across multiple pools to reduce risk of simultaneous interruption
Automatic Capacity Maintenance: Replaces interrupted or unhealthy instances to maintain target capacity [citation:2]
Health Check Integration: Can replace unhealthy instances automatically (for
maintaintype fleets) [citation:2]Hybrid Capacity: Mix Spot and On-Demand instances in same fleet for baseline capacity
Simplified Management: Single fleet request manages multiple instance types and Availability Zones
Elastic Scaling: Scale up/down by modifying target capacity without recreating the fleet [citation:3][citation:5]
Interruption Handling: Fleet can be configured to terminate, hibernate, or stop instances upon interruption