In this PR we improve the benchmarks reports to have them machine readable. The restoration benchmark is left out from this work as it's not going to appear in the CI-1 (gate-to-master) pipeline anyway as it takes 16 hours.
The involved benchmarks are
api
latency
memory
db
Each of them now creates a csv artifact during its new CI-1 step.
[x] Add a library to write csv reports of the benchmarks
[x] Use the lib in all benchmarks to get new csv artifacts in buildkite
[x] Add justfile runners for all benchmarks (-O0 is just for testing them)
[x] Add steps in the pipeline for all of them
[x] Add a library to parse pmap utility output
[x] Small rework of ToTextTracer and use it in place of heavier logging machine
In this PR we improve the benchmarks reports to have them machine readable. The restoration benchmark is left out from this work as it's not going to appear in the CI-1 (gate-to-master) pipeline anyway as it takes 16 hours. The involved benchmarks are
Each of them now creates a csv artifact during its new CI-1 step.
pmap
utility outputToTextTracer
and use it in place of heavier logging machineADP-3368
(this is a retake of https://github.com/cardano-foundation/cardano-wallet/pull/4621 that was closed automatically by github because of an operation on the base branch or smth)