support multiple benchmarks

[kwmonroe]

We should be able to list multiple benchmarks per unit defined in the test plan, ie:

bundle: bundle:~bigdata-dev/bigtop-processing-mapreduce-6
benchmark:
    resourcemanager/0:
        mrbench
        nnbench
        terasort
bundle_name: bigtop-processing-mapreduce
bundle_file: bundle.yaml

I would expect the output to add a graph for each benchmark.

[seman]

Currently it doesn't support multiple benchmarks. This is beta release.

[kwmonroe]

Yup, understood. Just logging it here as an issue / feature request.

FWIW, the code in run_actions() will already execute all benchmarks if they are enumerated as dicts, e.g.:

bundle: bundle:~bigdata-dev/bigtop-processing-mapreduce-6
benchmark:
    resourcemanager/0:
        mrbench:
            None
        nnbench:
            None
        terasort:
            None
bundle_name: bigtop-processing-mapreduce
bundle_file: bundle.yaml

So I think the only missing piece here would be generating graphs for each individual benchmark and writing those out to the results.html.