The requirements are vague saying that if you have a bigger load - you need more capacity - but what does that mean as a minimum like
RAM etc ..
can you please provide Atlassian suggestions on requirements?
Do you have any indications on your requirements? How many directories/users/groups/applications etc? Are you running a multi-node DC cluster? Do you expect N+1 redundancy?
We run a fairly significant Crowd environment (~ 70000 users, 12000 groups, 25 external directories, 40 applications) on 2 x c5.2xlarge AWS EC2 instances (8 vCPUs, 16GB RAM) on a 4GB heap, PSQL database running on it's own RDS instance and we can be running on a single node for maintenance/etc and the servers will barely break a sweat.