I'm trying to create some documentation for an AppStream project and need definitions for Fleet and Stack

Question

Unfortunately, AWS documents never actually define either Fleet or Stack, not (at least) in any way that I can use. I'd like someone who knows this technology intimately to provide their understanding of what a Fleet and a Stack are. I would particularly appreciate a definition that makes clever use of metaphor.

Answer 1

Fleet: Fleet is a collection of instances including Autoscaling group, AMI (images), instance type, and the base images (all need to define initially and can be edited). it combines everything in a single term known as the fleet. it can be either running mode or stop mode. it lets you update images. you create multiple fleets with the same definition or different means you can create multiple fleets with multiple requirements such as one for HPC, the second for high graphics, 3rd high memory use or storage, and the 4rth for normal purpose use. you can put all of them in running mode (you have to pay) or stop mode. it also has multiple types of instance starting modes (on-demand flexible,..etc)

Stack: In simple words, it will let you connect the fleet to the appstream access URL. meaning you can create a fleet but there is something needed that let you connect to the user pool and that is stack. this will let you connect the stack to the fleet which will be accessed through appstream URL. e.g. you have 3 fleets with all having a different configuration (HPC, high graphic, normal usage). so now either we create three stacks to connect it to 3 URLs or create one stack and associate it required fleets and now want to use 2nd just disassociate first one and associate 2nd fleet.

some cases: you need to update the fleet's image weekly. so you can create two fleets one for old and sec for new. where 2nd is associated with stack and getting all traffic) in update time you updated the first image (which takes 20-50 min to start or in running mode depending upon the image) while sec one is in already in production. when 1st is in running mode you replace it with 2nd one and stop the first one. OR if need to test the new upgrade one just attached it with the dev stack (new) and test, and after confirmation change it with the production one.