EC2: run_instances launches MinCount instances, MaxCount is ignored

Moto 5.1.18, boto3 1.34.162, Python 3.12

run_instances with MinCount=1, MaxCount=3 creates a single instance:

import boto3
from moto import mock_aws

@mock_aws
def repro():
    client = boto3.client("ec2", region_name="us-east-1")
    resp = client.run_instances(
        ImageId="ami-12c6146b", MinCount=1, MaxCount=3
    )
    print(len(resp["Instances"]))  # 1, expected 3

repro()

On real AWS the request launches as many instances as it can, between MinCount and MaxCount — so with capacity available this returns 3 instances, and fewer (but at least MinCount) only when capacity runs out. See "How instance launches work" in the RunInstances docs: https://docs.aws.amazon.com/AWSEC2/latest/APIReference/API_RunInstances.html

Moto only reads MinCount and uses it as the instance count, MaxCount is parsed but never looked at:

https://github.com/getmoto/moto/blob/master/moto/ec2/responses/instances.py#L49

min_count = int(self._get_param("MinCount", if_none="1"))
...
new_reservation = self.ec2_backend.run_instances(
    image_id, min_count, user_data, security_group_names, **kwargs
)

This goes unnoticed because pretty much everyone calls run_instances with MinCount == MaxCount (moto's own test suite does too, e.g. test_add_servers). We hit it with a batch launcher that intentionally uses MinCount=1, MaxCount=N and handles partial fulfillment — under moto every batch quietly degrades to a single instance, so the multi-instance code path was never actually tested.

Since moto has no capacity limits to hit, I'd expect it to launch MaxCount instances. Happy to send a PR if you agree that's the right behaviour.