I’ve been trying to figure out how capacity works in Bubble and how to predict capacity needed, but I have to say, everything is very vague from Bubble’s end on this. I am wondering how many units of capacity do you start with on the personal plan? Is it 1?
Also, does anyone have any experience with how many units were needed per x number of concurrent users? For example, if you have 100 users plan on having 20 units, 200 plan on 25 units, etc. This would be really helpful in understanding scalability of Bubble apps.
The short answer is it’s impossible to tell. Each app is different, so there are no rules of XX users = YY capacity units. Some folks are running fine with the baseline number of units, other apps need a ton of capacity to handle basic user load.
From my understanding, “capacity” doesn’t increase the maximum speed of your app, but it does increase the size of the pipe. Basically, adding capacity doesn’t make things faster, but it allows more of your users to work without tanking the system.
Again, the more important thing is what type of actions are the users going to be doing. Are there tons of API workflows? A lot of heavy data manipulation? Is it just basic CRUD actions? These things will all make a difference. Getting data from other users will be a good thing to have, but it may also be a red herring if your app isn’t comparable in terms of functionality.
I agree, but with the limited resources available I’m just hoping to get user data and make some assumptions off of it. Because again, what are tons of API workflows in Bubble terms? What is a lot of heavy data manipulation in Bubble terms? It’s all so vague that I’m hoping to just get some examples of Bubble users being able to handle many concurrent users. By large app I’m actually referring more to successful app - lots of users.
My last test couple of months ago showned it’s faster with 40 units (well the cost of it as to be justified). Before the new pricing, an API was taking under 200ms to respond and process, now the same api take 20 sec. With 40 units, it take 2 sec. Bubble is coming with (hopefully) on demand capacity, or I can be wrong. But behind the scene, node js and postgres are strong elements we can relies on. If you can run a lot of small api workflows, let say 20, it’s faster than running one workflow with 20 actions (a lot faster). Small processes are faster. Each plans have his own limits of simultaneous workflows you can run at once. You have to validate yourself to be sure (no official confirmation yet).
