In the effort towards moving {benchmarks} from using run_one() to run_benchmark() (and presumably eventually moving to a more abstracted orchestrator), the result JSON produced by {arrowbench} should contain all metadata necessary to post it to conbench—everything called context, info, tags, etc. on the {benchmarks} side.
This task is to rearrange as required such that a directory of JSONs produced by a run can be hoovered up and sent to conbench with no additional context or metadata.
In the effort towards moving {benchmarks} from using
run_one()torun_benchmark()(and presumably eventually moving to a more abstracted orchestrator), the result JSON produced by {arrowbench} should contain all metadata necessary to post it to conbench—everything calledcontext,info,tags, etc. on the {benchmarks} side.This task is to rearrange as required such that a directory of JSONs produced by a run can be hoovered up and sent to conbench with no additional context or metadata.