Benchmarking Suite

We have started thinking about various improvements we could invest in to make the framework more performant in a variety of ways, but when venturing down that path we would probably also be well served to have some minimal set of benchmarking tools that can be used to generate concrete data on how well a given node is performing. That will make it easy to both confirm success, identify problems and set concrete goals. Moreover, it will make it easier for third parties to figure out whether their infrastructure has the chops to take whatever load they have in mind in production.

It seems there are at least two ways one may want to benchmark for this purpose:

• Framework: Some general benchmark of the framework itself, possibly part of the overall testing suite.
• Instance: Its conceivable that performance may depend on the particulars of the blockchain and input schemas, as well as query statistics, hence, allowing someone to easily define and run benchmarks for a particular instantiated version of Hydra, e.g. like the Joystream query node, may of value.

I am not sure whether these genuine require distinct remedies, but the use cases are at least somewhat different.