Develop a robust server that receives GraphQL queries/mutations to record and analyze the traffic in multiple websites by multiple users.
Given that there are millions of users browsing millions of websites at any given minute, the server must be resilient, fault tolerant and able to handle lots of simultaneous requests.
This project must also include a load test client configured to post mutations to the server. The client doesn't necessarily need to be implemented in this project, since there are some other open source tools that can perform these queries.
It is encouraged that this project, as well as any potential database that it uses, is Dockerized
and runnable with docker-compose.
Every query and mutation must be covered by either unit or integration tests. The tests must be automated and reproductible.
The GraphQL schema defines the types, queries and mutations that need to be implemented to define the API.
Although the server can be implemented in any language (as long as it's exposed as a GraphQL endpoint), some languages are more encouraged than others and will yield 'bonus points' if used.
Encouraged languages:
- Go
- Elixir/Erlang
- Clojure
- Python
- JavaScript/TypeScript
- ... pretty much any functional language
The load test client that post mutations to the server must be configured in this project through a npm script, bash script, etc.
There are many libraries and packages that implement GraphQL clients, some suggestions are easy-GraphQL-load-tester and graphql-request