Skip to content

docs: Add histogram latency metrics documentation and test patterns#8675

Open
zbennett10 wants to merge 1 commit intotriton-inference-server:mainfrom
zbennett10:feat/histogram-latency-metrics-docs
Open

docs: Add histogram latency metrics documentation and test patterns#8675
zbennett10 wants to merge 1 commit intotriton-inference-server:mainfrom
zbennett10:feat/histogram-latency-metrics-docs

Conversation

@zbennett10
Copy link

Summary

Updates documentation and test patterns for the 5 new histogram latency metrics added in triton-inference-server/core#474.

Changes

docs/user_guide/metrics.md:

  • Add all 5 new histogram metrics to the Histograms table (request, queue, compute_input, compute_infer, compute_output)
  • Update the custom bucket boundaries section to list all 6 supported histogram families

qa/L0_metrics/metrics_config_test.py:

  • Expand INF_HISTOGRAM_PATTERNS from 1 entry to 6 entries to include the new histogram metrics

New Metrics Documented

Metric Description
nv_inference_request_duration_histogram_us End-to-end inference request duration
nv_inference_queue_duration_histogram_us Queue waiting time
nv_inference_compute_input_duration_histogram_us Input processing time
nv_inference_compute_infer_duration_histogram_us Model inference time
nv_inference_compute_output_duration_histogram_us Output processing time

Related Issues

Test Plan

  • Pre-commit hooks pass (black, flake8, codespell, license headers)
  • Copyright headers auto-updated by add-license hook
  • L0_metrics test suite passes with updated histogram patterns

Update metrics documentation and test patterns to include the 5 new
histogram latency metrics added in triton-inference-server/core.

- Document all 6 histogram metrics in the metrics user guide table
- List all histogram families that support custom bucket boundaries
- Update INF_HISTOGRAM_PATTERNS in metrics_config_test.py

Companion to triton-inference-server/core histogram latency metrics PR.

Addresses triton-inference-server#7672
@zbennett10
Copy link
Author

Companion PR -> triton-inference-server/core#474 😄

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Development

Successfully merging this pull request may close these issues.

1 participant