Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Thunderdome: Improve Experiment Quality #28

Closed
5 tasks done
iand opened this issue Aug 31, 2022 · 0 comments
Closed
5 tasks done

Thunderdome: Improve Experiment Quality #28

iand opened this issue Aug 31, 2022 · 0 comments
Labels
epic Overarching issue for an extended piece of work

Comments

@iand
Copy link
Collaborator

iand commented Aug 31, 2022

What Is It?

Provide more realistic measurment conditions and more options for fine tuning experiments to improve the quality of experiments.

Deliverables

  • New experiment option to warm up targets and adapt request load
  • New experiment option to filter incoming requests
  • New metrics to monitor performance of a target compared to live

Why Are We Doing It?

The initial development of Thunderdome fixed several factors that affect experiments for simplicity. For example requests are filtered to only allow send based requests starting with /ipfs or /ipns. Also, while Thunderdome probes targets to establish whether they are ready to accept requests, it immediately sends the full request load making metrics gathered at the start of the experiment skewed since the target has no chance to fill caches or grow queues.

We want to provide more options for users to control some of these factors so that experiments can be more representative of real environments. We also want to refine the way experiments report metrics by adapting the delivered load to available capacity and giving the test targets a chance to warm up into a stable configuration which will produce higher quality and more accurate metrics.

Notes

Tasks

@iand iand closed this as completed Sep 27, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
epic Overarching issue for an extended piece of work
Projects
None yet
Development

No branches or pull requests

1 participant