I want to create a Node.js API to retrieve feature flags data periodically (every 30 minutes). There is no database involved; the API returns hardcoded data. For 99% of the time, the same data is returned to all users. The monthly invocation of this API is around 500 million.
Before I invest time implementing this, could someone please share the estimated monthly cost?
It is not possible to provide a dollar amount nor recommended service tier for this request. 200 requests a second is reasonably voluminous, but it depends more on the complication level of the 1% of requests.
For all requests of this sort, you need to benchmark the application, and scale it up (whether in resources Starter → Standard → Pro, or number of concurrent instances) when/as necessary.
It depends on the language you use, the framework you use, the web server/worker process you use, the number of concurrent workers (threads/processes, separate from Render service instances) you use per instance, and it will be affected by the resource demands of the non-static response processing, that needs a suitable amount of headroom amongst the many other requests being handled at the same time.
There is no single formula for determining this, and it’s impossible for a service provider (Render) nor anyone without deep architectural knowledge of your application to make a reliable recommendation.
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.