How to scale Parse Server and live query to 30000 active users?

Parse Server 📡
#1

Hi all,

We are using Parse Server for our real time game and we will be growing in the upcoming months to at least 10.000 active users and maximum 30.000 active users.

We are using load balanced Parse Servers and Live Query Servers synced with Redit using GCP.

We would need some consulting services from any expert that could help us confirm that our infrastructure could manage this amount of active users and if not how could we design it.

Any expert available to have with us this consulting session?

Thanks so much

1 Like
#2

May be @davimacedo have some answers.

But in your use case, in a risk management perspective, i can suggest to simply try on a dedicated env 30 000 simultaneous live queries. With some scripts and tools, i think it’s feasible to simulate in real world this web socket connections. Many factors could play on your live query performance/limitations. I can also suggest may be take a look on how the livequery system is handled on parse server repo. The code is easily readable if you are familiar with javascript.

Side note: here the main bottleneck will be the Redis database. If redis is slow and fail to scale it will be the main issue. another point is parse server memory usage, how many memory usage take your live query subscriptions on server side. You can may be check your memory usage based on your current user base.

#3

Unfortunately I can’t offer consulting services myself but load balancer on top of many parse server instances using the same redis instance is the way to go. There is a small section talking about that on Parse Server guide: Parse Server Guide | Parse

#4

@realsail Were you able scale well or ran into issues?

#5

I will work with the Load Balancer and Redis for the Live Query Servers and test it.

I will share with you the results.

Thanks so much to all for all your feedback.

I really appreciate it

#6

Did you get any good results?