auto scaling
Hello Railway team. In 2 weekends I expect there to be significantly more traffic on my application than normal, I'm trying to do everything I can in advance to make sure our servers can handle the demand.
The weekend I'm concerned about will experience about 4x the demand that we've serviced until now.
Our software manages competitive wrestling tournaments (youth, high school, and college), and there are typically a few hundred competitors and their families present at each event. We cannot afford to hit scaling issues that would lock us up, as users don't have the luxury of coming back later (a laggy site would be ok, but if it didn't work at all we would be screwed).
My mentor told me to look into autoscaling, to make sure that my provider doesn't limit our service. Can you give me any advice or suggestions on how to make sure we are ready? Thank you.
7 Replies
Project ID:
1fc9ac5f-29f8-4b23-ab28-1eebdaeb8c061fc9ac5f-29f8-4b23-ab28-1eebdaeb8c06
We also have our frontend here: 96035335-03de-4dc8-9ee2-2271eacad269
Railway should be able to handle it as long as your app can ( and assuming you don't hit the limit of pro plan )
Railway autoscales up until the max allowed for your plan
which in this case is pro plan
How can I tell what the limit is? Or if I am approaching it?
Pro plan has a limit of 32 vCPU per service and 32 GB RAM per service ( and 50 GB volume storage )
You can see whether you're approaching the limit or not in your account usage page or service metrics page ( recommended )
You can see your service metrics by going to your project, clicking on a service, and then click on the
Metrics tabIs there a way to see historical data of account usage? I can only see the past week, I'd like to compare to our previous periods of heavier usage.
Also, forgive me if I don't understand the terminology, but is my volume storage the egress? Thank you.
I'm not sure if you can see past metrics, I think if you want to do that you'd need to export your metrics to your own server and store them there
I could be wrong though
volume storage is just persistent storage on railway
you can attach a volume to a service and specify the filepath of the mount, and anything stored in that filepath will persist between deploys