Thanks for the info, though i still

[fringe raptor]

You can view spending over time on the Dashboard, I don't think there are documented, public APIs for this, but that's an interesting idea that maybe we could surface them and make them official - I'll see what people think about that.

I think the reality is the free tier thresholds are so high by that the time a game has the sorts of player activity levels that it goes beyond what's included for free, that it's not a problem for developers or studios.

It's going to be different across services and different games as they will all use services differently, but for a product like Cloud Save, if you have less than 50,000 monthly active players then with typical usage (1 million read/writes + storage) it can reasonably be expected to work out as free to use for around the first 50,0000 active players.

Even if you 10x those player numbers and find you have a smash hit on your hands, with the sort of player numbers of a chart topping game like Helldivers 2, then your Cloud Save bill for a game with the same usage pattern as above might still be only $50-60 a month.

Of course your cost per player comes down to exactly how a game uses the service, the nature of some games is that some folks are going to find it costs them nothing to run a game with anywhere from 25,000-100,000 monthly active players using services like Cloud Save, Cloud Code, Leaderboards, etc (and of course including use of services like Authentication, which we don't explicitly charge for).

[past bramble]

I think thats a valid point, thank you very much for taking the time to write this out. And yeah as i also mentioned i am fairly sure safe guards are good enough that i will most likely not run into this issue. And my main concern was with lobbies/relay/matchmaking/game hosting. As i think those can get more expensive, but i do see theres a way to limit max game servee count as i understand. So that should help

[fringe raptor]

Where I think it maybe gets tricky for developers, especially folks not in a larger org where people have experience in designing architecture for multiplayer games (and someone else on the hook for the bill), is that they could end up using services in an expensive way.

e.g. for example by using Multiplay run dedicated game services for things that using Relay might be a better fit for (because it's less expensive) or just by writing code that works in an unintended way and releasing something that calls a service way more than it should.

..

Yeah I think that example you just gave of game server hosting is specifically a tricky one.

There are some awesome things you can only really do with a dedicated game server, but it's also much more expensive and agree getting a bill for that is the sort of thing that would keep me up at night if I was personally on the hook for the bill!

One approach I think is super interesting is the idea of combining things like Relay (Peer 2 Peer networking) with things like Cloud Code to handle server authoritative logic so that, unless you have something that really needs a dedicated game server (eg a competitive twitch shooter), you can build a game with multiplayer experiences but with server authority for things that need it (eg loot drops).

While this is something developers can do today, making that easier to do - without developers having to spend loads of time trying to work out how to use all the things together and what the implications are for their game - is something we discuss internally and would like to solve for.

[past bramble]

Yeah i think that approach is great for card games or turn based games indeed. I am trying to build more realtime games so that requires a more direct connection, and client hosted is not an option as i dont want to allow the host to cheat. So that actually left game servers as only option which was why i was worried about the safeguards. Especially a problem where most multiplayer games nowadays need to start with a f2p architecture to garner enough audience to maintain proper matches. So all these constraints made it quite worrisome. When looking at azure cloud services i do see that they provide a more vast capability to take actions based on spending, which is cool, but then its outside of unity eco system and less comfortable

[fringe raptor]

That is really good feedback, thank you.

Both you asking about an API, and thinking about how we recently added Cloud Code Triggers (to Cloud Save, Leaderboards, etc), makes me wonder if we could do something similar there to help folks manage costs, especially with services that could potentially get expensive.

I'm out the rest of this week, but I'll bring that up with people when I am back next week.