-
Notifications
You must be signed in to change notification settings - Fork 56
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
Extend Common Benchmark Fields #296
Comments
This would be fantastic! It would also allow us to create really nice and specific templates for elasticsearch. One question that I have, which @mfleader talked to me about, was about the definition for
I think that the metadata field is a bit confusing. We might want to look at the current fields we are placing there and whether or not they should be added into |
yeah metadata is a bit weird. I'm fine if some of the fields I mentioned (start_time, end_time, run_id, etc.) are in that metadata class in the actual code as long as it gets flattened in the doc. cluster_name shouldn't be there though since it's environment specific |
For newly proposed benchmark wrappers, let's say I have a field in my |
How is |
Common Benchmark Fields
With @learnitall's PR merged we have a launch pad to start creating a well defined schema for our benchmarks. The biggest gap I've seen is around common fields within benchmark results.
This proposal is to create strictly defined fields across all benchmarks, regardless of benchmark/environment/config. This data should consist of:
This would look something like this:
These fields should take precedence over any field defined with the same key in
metadata
/config
/data
/labels
. For instance if benchmark creates a field calledstart_time
in it's data dictionary, then it should be overwritten by the top-levelstart_time
field.Common logic for setting the fields
Subsequently, the
Benchmark
Abstract class should enforce the definition of these fields while providing users the ability to override if required.Psuedo-code Example:
Ideally we can create a common run method that ensures generated fields like
start_time
andend_time
get added, but it'll require tweaking how the Benchmark class works atm.The text was updated successfully, but these errors were encountered: