QN/Hydra: Low-hanging load testing

[bedeho]

Background

With Orionv2, both QN and Orionv1 will not longer be relevant for read performance in consumer apps, but since it won't be ready for some time, it may be worth profiling the capacity of those nodes to see if there are issues at whatever scale we have in mind, and what low-hanging improvements we can do that are worth doing.

Proposal

1. Propose capacity requirements for QN
2. Propose a capacity benchmarking plan for QN
3. Execute.
4. Report results and range of remedies

[traumschule]

not exactly in scope but sync performance would increase dramatically with #3748
does this ticket only target hydra or are you also interested in comparison to potential Subsquid implementations? #2893

• Orion integration tests proposal orion#147