Python: Fix workflow runner concurrent processing

[TaoChenOSU]

Motivation and Context

Currently, the runner delivers messages from a source executor one after another regardless of the target. This means if a source sends more than one messages to multiple targets, the targets will be receiving those messages one after another. For example, executor A sends message X to executor B and message Y to executor C, message Y will not be delivered to executor C until executor B finishes processing message X. This breaks our concurrency promise.

Description

The runner will now deliver messages to all targets concurrently.

Summary of the current behavior:

• A single message from a source to multiple targets will be delivered concurrently to all targets.
• Multiple messages from a source to the same target will be delivered in the order they are sent.
• Multiple messages from different sources to the same target will be delivered to the target one at a time without order guarantee because Python doesn't have true parallelism.
• Multiple messages from different sources to different targets will be delivered concurrently to all targets, assuming each message is targeting a unique target, or it falls back to the previous scenario.

Contribution Checklist

• The code builds clean without any errors or warnings
• The PR follows the Contribution Guidelines
• All unit tests pass, and I have added new tests where possible
• Is this a breaking change? If yes, add "[BREAKING]" prefix to the title of the PR.

[markwallace-microsoft]

Python Test Coverage Report •

File	Stmts	Miss	Cover	Missing
packages/core/agent_framework/_workflows
_runner.py	173	2	98%	304–305
TOTAL	22151	3472	84%

Python Unit Test Overview

Tests	Skipped	Failures	Errors	Time
4264	240 💤	0 ❌	0 🔥	1m 16s ⏱️

[Copilot]

Pull request overview

This PR updates the Python workflow Runner’s message-delivery loop to improve concurrency so that messages from a single source can be delivered to multiple downstream targets in parallel (instead of being blocked behind one another).

Changes:

• Refactors Runner._run_iteration() to parallelize message delivery across edge runners while preserving per-edge-runner message ordering.
• Adds unit tests asserting (1) per-edge-runner ordering and (2) concurrent execution across multiple edge runners for the same source.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 4 comments.

File	Description
python/packages/core/agent_framework/_workflows/_runner.py	Changes _run_iteration() delivery scheduling to run different edge runners concurrently while keeping ordering within each edge runner.
python/packages/core/tests/workflow/test_runner.py	Adds tests for per-edge-runner ordering and concurrency across multiple edge runners.