I've created this repo as a learning opportunity to try/learn different approaches to do requests in batch as fast as possible keeping the code simple and readable.
The idea is to simulate a scenario where an application has to do multiple HTTP requests to an specific API retrieving the result as fast as possible
These options below will be used for this first version
- Sequential blocking requests
- Parallel requests
- with CompletableFuture
- with ParallelStream
- with WebFlux/Reactor
- with CoRoutines
Collect those metrics to have a better understanding of the application results
- Measure application execution time
- Identify CPU/Memory utilization
The repo contain two modules, the webapp that should always be running, so the batch-requester module can crete the requests.
This module is a simple SpringBoot application using webflux to allocate a low amount of resources.
There are two endpoints:
/millis-> that will respond with a random delay between 100ms to 500ms/seconds-> that will respond with a random delay between 1s to 5s
Both endpoints receive an optional RequestParam that will be included in the String used in its response so it can be used to tracked which request has been received.
This module is using OpenFeign to create the http client to realize the requests to the webapp.
Even though, its a SpringBoot application, it is not going to start as a web application and the application will stop after the requests are done
Also to be able to measure the resources usage there is a wait for input before the application starts