Add model latency endpoint

[A-Artemis]

Summary

Added a MetricsService which collects and retrieves the model's inference latencies. It uses shared memory so that it is process safe when accessing the recorded measurements.
The latencies are stored in memory and only a total of 1024 entries are kept.

How to test

http://[geti-tune.localhost/api/pipelines/ace3f1da-fdd9-4048-a95e-a647ed969442/metrics](http://geti-tune.localhost/api/pipelines/ace3f1da-fdd9-4048-a95e-a647ed969442/metrics)

Checklist

• I have added unit tests to cover my changes.​
• I have added integration tests to cover my changes.​

License

• I submit my code changes under the same Apache License that covers the project.
  Feel free to contact the maintainers if that's a concern.
• I have updated the license header for each file (see an example below).

# Copyright (C) 2025 Intel Corporation
# SPDX-License-Identifier: Apache-2.0

[github-actions]

Docker Image Sizes

Image	Size
geti-tune-backend-pr-4599	1.2G
geti-tune-backend-sha-fc057a8	1.2G
geti-tune-ui-pr-4599	50M
geti-tune-ui-sha-fc057a8	50M

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[Copilot]

Pull Request Overview

This PR adds model latency tracking capabilities by implementing a MetricsCollector that uses shared memory for process-safe storage of inference latency measurements, along with a new API endpoint to retrieve pipeline metrics.

• Implements a MetricsCollector using shared memory to track model inference latencies across processes
• Adds pipeline metrics API endpoint that returns latency statistics (avg, min, max, p95, latest) over configurable time windows
• Integrates latency collection into the inference workflow by recording start/end times and storing measurements

Reviewed Changes

Copilot reviewed 12 out of 12 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
backend/app/services/metrics_collector.py	Core MetricsCollector implementation with shared memory and circular buffer
backend/app/workers/inference.py	Integration of latency measurement recording in inference workflow
backend/app/services/pipeline_service.py	Pipeline metrics calculation and percentile computation logic
backend/app/api/endpoints/pipelines.py	New GET endpoint for retrieving pipeline metrics with validation
backend/app/schemas/metrics.py	Pydantic models for metrics API response structure
backend/app/services/model_service.py	Enhanced LoadedModel to include model ID for metrics tracking
backend/app/schemas/model_activation.py	Added active_model_id field to ModelActivationState
backend/tests/unit/services/test_metrics_collector.py	Comprehensive unit tests for MetricsCollector functionality
backend/tests/unit/services/test_pipeline_service.py	Unit tests for pipeline metrics calculation

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[itallix]

Great progress! I've added a few suggestions for further improvement.

[itallix]

LGTM!