feat(benchmarking): enhance inference-perf-bench tooling - #540
Open
alizaidis wants to merge 1 commit into
Open
Conversation
alizaidis
force-pushed
the
feat/inference-benchmarking
branch
from
August 7, 2026 19:25
2690528 to
93f4418
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Description
This PR enhances our implementation of
inference-perf-benchload-testing utility to be more robust, flexible, and resilient to missing configuration values.Key Changes
configure_benchmark.shto dynamically deriveAPP_LABEL,INFERENCE_BASE_URL, andBENCHMARK_STAGE_DURATIONif they aren't provided in the environment. This prevents structurally invalid Kustomize templates from being generated during CI or local testing.run_benchmark.shso that if a model-specific configuration directory doesn't exist for the current model, the script will gracefully fall back to thebasebenchmark configuration directory rather than failing out.secretproviderclass-huggingface-tokens.tpl.yamlto the benchmark template generation flow to ensure the benchmarking jobs can mount tokens necessary for testing gated models.benchmarking.tpl.envdynamically interpolates the proper target namespaces and service accounts instead of relying on hardcoded environment names.Motivation and Context
These utility improvements make the benchmarking suite significantly easier to use when testing new or custom models. By handling missing environment variables and missing model-specific directories natively, it removes manual boilerplate required from developers attempting to load-test their deployments.
Type of Change