feat(copaw): propagate traces to child agents and suppress duplicate entry spans#164
Open
Cirilla-zmh wants to merge 2 commits intoalibaba:mainfrom
Open
feat(copaw): propagate traces to child agents and suppress duplicate entry spans#164Cirilla-zmh wants to merge 2 commits intoalibaba:mainfrom
Cirilla-zmh wants to merge 2 commits intoalibaba:mainfrom
Conversation
Change-Id: Iadc503014273de1c9455ff6f7518474083dbf1a8 Co-developed-by: Cursor <noreply@cursor.com>
Cirilla-zmh
added
enhancement
Change-Id: Ibfc5a4ad46670bd1bb18bcefcad95f876ebf389d Co-developed-by: Cursor <noreply@cursor.com>
Contributor
There was a problem hiding this comment.
Pull request overview
This PR enhances the CoPaw OpenTelemetry instrumentation to keep a single trace across parent/child CoPaw agent processes invoked via AgentScope shell tooling, while avoiding duplicate “entry” spans in child processes.
Changes:
- Injects current trace context into
execute_shell_commandsubprocess environments forcopaw agents chat(or via opt-in env flag). - Suppresses
enter_ai_application_systementry span creation in child CoPaw processes and attaches extracted parent context instead. - Adds supporting modules, tests validating propagation/suppression behavior, and updates documentation + changelog.
Reviewed changes
Copilot reviewed 9 out of 9 changed files in this pull request and generated 4 comments.
Show a summary per file
| File | Description |
|---|---|
| instrumentation-loongsuite/loongsuite-instrumentation-copaw/src/opentelemetry/instrumentation/copaw/_shell_patch.py | Wraps AgentScope shell execution to inject trace env and mark child processes. |
| instrumentation-loongsuite/loongsuite-instrumentation-copaw/src/opentelemetry/instrumentation/copaw/patch.py | Updates query_handler wrapper to suppress entry in child mode and attach extracted context. |
| instrumentation-loongsuite/loongsuite-instrumentation-copaw/src/opentelemetry/instrumentation/copaw/_env_carrier.py | Adds env getter/setter carrier for propagator injection/extraction. |
| instrumentation-loongsuite/loongsuite-instrumentation-copaw/src/opentelemetry/instrumentation/copaw/_constants.py | Centralizes env var names and child-process detection helper. |
| instrumentation-loongsuite/loongsuite-instrumentation-copaw/src/opentelemetry/instrumentation/copaw/init.py | Instruments/uninstruments AgentScope execute_shell_command in addition to query_handler. |
| instrumentation-loongsuite/loongsuite-instrumentation-copaw/tests/test_shell_propagate.py | Tests shell command matching and env injection of TRACEPARENT + child marker. |
| instrumentation-loongsuite/loongsuite-instrumentation-copaw/tests/test_child_entry_suppression.py | Tests child mode suppresses entry span emission. |
| instrumentation-loongsuite/loongsuite-instrumentation-copaw/README.md | Documents sub-agent CLI trace continuity and the opt-in forced injection flag. |
| instrumentation-loongsuite/loongsuite-instrumentation-copaw/CHANGELOG.md | Notes the multi-agent trace propagation and child entry suppression feature. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
+68
to
+69
| await asyncio.wait_for(proc.wait(), timeout=timeout) | ||
| stdout, stderr = await proc.communicate() |
Comment on lines
+147
to
+157
| env = _build_subprocess_env() | ||
| try: | ||
| return await _run_shell_command_with_env(command, timeout, env) | ||
| except Exception: | ||
| logger.debug( | ||
| "%s.%s inject path failed; falling back to original", | ||
| _MODULE_SHELL, | ||
| _PATCH_TARGET, | ||
| exc_info=True, | ||
| ) | ||
| return await wrapped(*args, **kwargs) |
Comment on lines
+80
to
+91
| try: | ||
| proc.terminate() | ||
| stdout, stderr = await proc.communicate() | ||
| stdout_str = stdout.decode("utf-8") | ||
| stderr_str = stderr.decode("utf-8") | ||
| if stderr_str: | ||
| stderr_str += f"\n{stderr_suffix}" | ||
| else: | ||
| stderr_str = stderr_suffix | ||
| except ProcessLookupError: | ||
| stdout_str = "" | ||
| stderr_str = stderr_suffix |
| ) | ||
| except Exception: | ||
| logger.debug("Failed to inject trace into env", exc_info=True) | ||
| return merged |
|
This PR has been automatically marked as stale because it has not had any activity for 14 days. It will be closed if no further activity occurs within 14 days of this comment. |
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
What changed
Shell subprocess: trace propagation (
multi_agent_collaboration)When AgentScope’s
execute_shell_commandruns a command that looks like a CoPaw sub-agent chat (copaw+agents+chatin the command string), the instrumentation merges the current trace context into the subprocessenvusing W3CTRACEPARENT/TRACESTATE(and configured propagators), aligned with OpenTelemetry’s environment carrier semantics. It setsCOPAW_OTEL_CHILD_AGENT=1so the child recognizes the role. An opt-inCOPAW_OTEL_INJECT_SHELL_TRACEforces injection for every shell invocation (documented as advanced / risky for non-CoPaw children).Entry span behavior in child processes
AgentRunner.query_handleris updated so that whenCOPAW_OTEL_CHILD_AGENTindicates a child agent process, it does not create a newenter_ai_application_systemspan; it attaches context extracted from the environment so AgentScope and other spans continue in the same trace as the parent.Supporting modules
_env_carrier.py— local environment getter/setter for propagators (mirrors upstream_envcarrierwhere the published wheel may not expose it)._constants.py— names and helpers forCOPAW_OTEL_CHILD_AGENT, inject flags, and child-process detection._shell_patch.py— wrapsexecute_shell_commandto merge env and delegate to the same async shell behavior with explicitenv=.Tests
instrumentation-loongsuite/loongsuite-instrumentation-copaw/tests/test_shell_propagate.py— asserts trace env injection andCOPAW_OTEL_CHILD_AGENTfor matching shell commands (and related behavior).instrumentation-loongsuite/loongsuite-instrumentation-copaw/tests/test_child_entry_suppression.py— asserts child mode suppresses entry creation while attaching propagated context.Documentation
instrumentation-loongsuite/loongsuite-instrumentation-copaw/README.md— new section Sub-agent CLI and trace continuity (multi_agent_collaboration), including baggage /OTEL_PROPAGATORSnotes and advancedCOPAW_OTEL_INJECT_SHELL_TRACEwarning.Type of change
Please delete options that are not relevant.
How Has This Been Tested?
Please describe the tests that you ran to verify your changes. Provide instructions so we can reproduce. Please also list any relevant details for your test configuration
Does This PR Require a Core Repo Change?
Checklist:
See contributing.md for styleguide, changelog guidelines, and more.