[struct_t]

"a certain amount of times as power gets divided by how many users are using it."
Then don't limit userbase, change functionality. Slot attacks through peak attack times. Like, use a backlog and limit the amount of synchronous attacks.

[void account]

Why not take it the extra step and simply set up servers for people? I remember a long time ago on a different forum I sold "Help-You" packages which I guided newbies with some $ to spend into making their own servers. I gave them a few scripts too... Won't be doing that here though, I'm definitely not caught up with today's stresser competition and don't plan on getting back in on it.

Just a thought, OP.

[void account]

I could do that, but I think long term wise the stresser is better

What prices do you think I should do it on a month?

You should use a simple algorithm that calculates how many users the big boys in the Stresser market have, then correlate that with the power and price, and simply plug and play with numbers to have a sustainable model. Make sure you factor in static costs such as server upkeep, whatever you value your time to be, website protection, legal documentation if you plan to make this bigger, and so on. I'm not sure how far you want to go so I'll just stop here.

[struct_t]

This.. some customers dont need instant response. Just queue a strong attack 4h beforehand for 3600 secs rather than a weak one instantly. Im amazed nobody has a queue based source in use.

Yeah, I disagree about the 4 hour before hand thing. I think a backlog queue system with a limit of how many attacks threads can be running is important.