Apologies on the size of this PR as lots of separate workstreams were handled and I neglected to merge back here. Lots of features added on this one.
Panda reports! Canned ASV reports haven't been working so well for throughput so I created a custom report in the form on a Jupyter notebook using Pandas to parse the JSON output and plot some metrics.
Docs. Started more extensive docs on a Jupyter notebook.
Improvements to Dask stability. Made various tweaks and fixes to increase stability of these tests. Mostly related to chunk configuration.
Standardize parameters. Made parameters uniform throughout all tests so that parsing JSON output to generate reports is easier to do.
KubeCluster wait function feature. Added feature to simply move forward with test as long as a threshold is met. Currently hard-coded to 5 minute wait and workers is 95% of requested config (e.g., 100 workers is requested, but stuck on 95 or more for more than 5 minutes, the test will just run). Next PR will include some sort of feedback saying the results may be tainted because of this.
Apologies on the size of this PR as lots of separate workstreams were handled and I neglected to merge back here. Lots of features added on this one.