Use a Poison distribution. This creates a vector of 3600 ints each of which has the number of customers that came in during that second. The vector's index is the timestamp.
Keep in mind that there will be times when more than 1 customer comes in during any given second. If so each customer would have the same timestamp. Change the time resolution as desired. You can also just create a vector of tuples with time and customer_count and push an entry only one or more customers arrive at any given second. That's your assignment. This is to illustrate a Poisson distribution.
#include <iostream>
#include <random>
#include <vector>
int main()
{
std::random_device rd;
std::mt19937 gen(rd());
std::poisson_distribution<> d(5/60.0); // set average for Poisson distribution for 5 per minute
std::vector<int> customers_at_each_second;
int total_customers = 0;
for (int i = 0; i < 3600; i++)
{
int customers_at_this_second = d(gen);
customers_at_each_second.push_back(customers_at_this_second);
total_customers += customers_at_this_second;
}
std::cout << "Total Customers in the hour " << total_customers << '\n';
}