Skip to content

[Feature Request] Measure request dispatch accuracy #110

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

Closed
markVaykhansky opened this issue Apr 8, 2025 · 1 comment
Closed

[Feature Request] Measure request dispatch accuracy #110

markVaykhansky opened this issue Apr 8, 2025 · 1 comment

Comments

@markVaykhansky
Copy link
Collaborator

Description
When using constant or poisson rates we would like to know how accurate the dispatch of requests really is.
So for each request we would like to have scheduled_time and actual sent_time as well as part of a summarized report file.

For example, if we have a constant rate of 1 RPS it would be meaningful to know that a request was sent after 1.2 seconds instead of 1 second.

Motivation
When analyzing the benchmarks in order to make decision regarding production grade workloads we need to know how accurate the measurement was. If the data is very skewed it would be wrong to make decisions based on it. Such problems are expected to arise under high load which makes it even more crucial to investigate as this is where systems typically fail.

@markurtz
Copy link
Member

Closing this out now that #96 has landed on main. Each request stored in the output now contains a "scheduler_info" object which contains the request times here along with any other important times for the benchmarking process.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants