Reporting of Generated Benchmarks
We require the benchmark numbers from the marconi-bench tool to be correlated and graphed. We could do this by running the benchmark 'n' times, increasing load by a factor every time and have numbers generated for latency, throughput. This could then be graphed over time, to see the trends.
To run this automatically, it would require a benchmarking cluster to be setup which is referenced in https:/
That being said, first step is just to be able to generate a graph (gnuplot?) for a single run (containing multiple iterations with progressively higher load, as explained above).
NOTE: Once we have initial results, publish on wiki and/or the mailing list. We should also demo in Paris summit.
Blueprint information
- Status:
- Not started
- Approver:
- Flavio Percoco
- Priority:
- Medium
- Drafter:
- Sriram
- Direction:
- Approved
- Assignee:
- Nataliia Uvarova
- Definition:
- Approved
- Series goal:
- Accepted for kilo
- Implementation:
-
Not started
- Milestone target:
- None
- Started by
- Completed by
Related branches
Related bugs
Sprints
Whiteboard
let's for the first patch do the "simplest thing that could possibly work" (kgriffs)
Keeping it for now. I wonder whether this blueprint still makes sense if we rely on Rally for benchmarks.
-- flaper87 2014-08-4
Gerrit topic: https:/
Addressed by: https:/
Improve benchmarking tool
Work Items
Dependency tree
* Blueprints in grey have been implemented.
