ec2 throttling and batch operations

I'd like to understand how batch operation against ec2 count toward throttle limits. Two requests specifically are of interest to me:

• LaunchFleet using EC2Fleet configuration -- if I'm attempting to start 2,000 instances with a single LaunchFleet request and also apply a tag to each instance on launch (as part of launch fleet configuration), how does this count against the throttle limits? Will only 1,000 be started and then I should wait until the bucket is refilled? What about the tagging of instances on launch, does this constrain things further?
• TerminateInstances -- similar question. I know that the max number of instances I can terminate in one call is 1,000 but if I make one such call, does the bucket empty and I should wait 1 minute before terminating another batch?

thanks