Open kckeiks opened 1 year ago
Could you elaborate on this?
Could you elaborate on this?
👍 I updated the title and added a description.
Can we create a bench
for each crate we want to benchmark? And then just add one command to ci to run all the benches
I don't have any particular preference atm. I imagine that whoever picks this work can assess. I think the minimum requirement should be that the whole flow (querying indexer, requesting from node) is tested.
I don't have any particular preference atm. I imagine that whoever picks this work can assess. I think the minimum requirement should be that the whole flow (querying indexer, requesting from node) is tested.
not about preference, whether we decide to do a full benchmark crate to do all of this or are we benchmarking a specific crate.
It would be useful to be able to run a stress test for the gateway on command.
Suggestion:
Consider that we should run all participants (indexer, node, gateway) locally in test so that latency (due to its volatility) does not affect the results.