a year ago
I need to run a backend service that supports 66 requests per second, totaling 178 million requests per month. Which plan should I choose? The backend service is running an MQTT broker and accepting MQTT requests. Additionally, I have a simple frontend Next.js and backend Nest.js application for a monitoring web app.
16 Replies
a year ago
Hobby plan and Pro plan is more about what your service's purposes are, this seems like a commercial project - in which case you'd want to take pro plan. Hobby plan is only meant for non-commercial purposes
Yes, but there is a budget concern. These IoT devices log data to a service.
Got it. But how can I understand what 32 GB RAM / 32 vCPU per service means? Is there any reference material available for this?
a year ago
Of course! This might be helpful:
32 vCPU and 32GB RAM per service means that each service you run on Railway can scale up to 32vCPU or 32GB Ram
on Hobby plan you can only scale up to 8vCPU and 8GB Ram per service
Ah, I see. In this plan, I can manage up to those cores. When I want to deploy something, I'll need to choose a plan based on the resources I require. right ?
a year ago
When you select a plan then you're given an account under that plan and you can deploy as many things under than plan as you like, so you won't have to create a new pro plan or hobby plan for each service you deploy, you can have them all under one plan
Okay, so when I choose the Pro plan, I’ll need to pay separately for resources like CPU, RAM, and network, right
a year ago
Yes, you pay 20 dollars per person on the plan and then you're also charged for your total resource usage per month
a year ago
^ For pro plan
thanks for info, one more question how can I determine the usage cost for the traffic I described earlier?
a year ago
It's hard to say since it's very dependant on your service, for example 178 million requests per month for a static JSON response wouldn't cost as much as 178 million requests per month for big documents or information that requires a lot of processing
You'd need to calculate this yourself based on your current or predicted resource usage of your app and calculating the cost based on
If your app is already deployed and you have metrics for your app's resource usage then it shouldn't be hard to calculate, get your base CPU usage, memory usage, disk usage and network egress and calculate the cost of them based on railway's pricing before adding it together, that should give you an idea of how much your app will cost per month running on Railway
a year ago
You can use this calculator Railway provides on their pricing page to help you
a year ago
No problem at all <:SmileFluent:865334941150871592>