Add session-aware evaluation logging

[AQing-527]

This pull request enhances logging and traceability across the evaluation and experience retrieval modules by consistently including session identifiers in log messages. It also introduces a new utility class to standardize session ID extraction and enriches context propagation for experience queries.

Logging and traceability improvements:

• Added a new utility class EvaluationLogContextHelper to extract session IDs from EvaluationContext and RunnableConfig, and refactored code to use this helper for consistent session ID retrieval across modules.
• Updated log statements in CriterionEvaluationAction, EvaluationObservationLifecycleListener, and ExperienceRetrievalEvaluatorFactory to include sessionId for better traceability of evaluation and experience retrieval operations. [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11]

Evaluation and experience context propagation:

• Enhanced the experience retrieval process to propagate additional context (sessionId, tenantId, userId) from CriterionExecutionContext to ExperienceQueryContext, improving downstream query traceability and multi-tenancy support. [1] [2]

Minor refactoring and cleanup:

• Added missing imports and performed minor code cleanups in affected files. [1] [2] [3]

These changes collectively improve observability, debugging, and context-awareness in the evaluation pipeline.

[canfuu]

LGTM

[Copilot]

Pull request overview

This PR improves observability across the evaluation and experience-retrieval pipeline by standardizing session ID extraction and including sessionId in key log messages, and by propagating more identity context into experience queries.

Changes:

• Introduces EvaluationLogContextHelper to consistently extract sessionId from EvaluationContext and RunnableConfig.
• Updates evaluation lifecycle/executor logging to include sessionId for better traceability.
• Enriches experience retrieval by propagating sessionId/tenantId/userId into ExperienceQueryContext.

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
assistant-agent-extensions/src/main/java/com/alibaba/assistant/agent/extension/experience/model/ExperienceQueryContext.java	Adds sessionId and userId fields to support richer downstream experience-query context.
assistant-agent-extensions/src/main/java/com/alibaba/assistant/agent/extension/evaluation/hook/BeforeAgentEvaluationHook.java	Adds session-aware logs during evaluation start/context build/completion and threads context into result logging.
assistant-agent-extensions/src/main/java/com/alibaba/assistant/agent/extension/evaluation/experience/ExperienceRetrievalEvaluatorFactory.java	Logs session-aware experience retrieval activity and propagates session/tenant/user identifiers into experience queries.
assistant-agent-evaluation/src/main/java/com/alibaba/assistant/agent/evaluation/util/EvaluationLogContextHelper.java	New utility for consistent session ID extraction (context env + config thread id).
assistant-agent-evaluation/src/main/java/com/alibaba/assistant/agent/evaluation/observation/EvaluationObservationLifecycleListener.java	Adds session-aware lifecycle logging for graph start/end/errors and criterion execution boundaries.
assistant-agent-evaluation/src/main/java/com/alibaba/assistant/agent/evaluation/executor/CriterionEvaluationAction.java	Adds session-aware logging for criterion execution, batching behavior, and errors.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The log message labels the last field as result={}, but the code passes result.getValue(). This is confusing when debugging; either log the full CriterionResult (if safe) or rename the label to value={}.

Suggested change

	logger.info("Criterion {} completed, sessionId={}, status={}, result={}",
	logger.info("Criterion {} completed, sessionId={}, status={}, value={}",

[Copilot]

New behavior propagates sessionId/tenantId/userId from CriterionExecutionContext into ExperienceQueryContext, but there’s no test asserting that these fields are passed through to ExperienceProvider.query(...). Since this module already has unit tests around ExperienceQueryContext, consider adding a focused test that mocks ExperienceProvider and verifies the constructed ExperienceQueryContext contains the expected IDs.

[Copilot]

queryText appears to be derived from end-user input, but it’s currently logged at INFO level. This can leak PII/secrets into logs. Consider removing it, logging only its length / a truncated value, or downgrading to DEBUG with truncation/redaction.

Suggested change

	log.info("ExperienceRetrievalEvaluatorFactory#{} - reason=开始检索经验, sessionId={}, queryText={}, experienceTypes={}",
	evaluatorId, sessionId, queryText, experienceTypes);
	int queryTextLength = StringUtils.hasLength(queryText) ? queryText.length() : 0;
	log.info("ExperienceRetrievalEvaluatorFactory#{} - reason=开始检索经验, sessionId={}, queryTextLength={}, experienceTypes={}",
	evaluatorId, sessionId, queryTextLength, experienceTypes);

[Copilot]

Session ID propagation into ExperienceQueryContext only checks environment.sessionId, but EvaluationLogContextHelper also falls back to environment.threadId. This can cause logs to show a non-null sessionId while downstream experience queries receive a null sessionId. Consider using EvaluationLogContextHelper.getSessionId(executionContext.getInputContext()) (or duplicating the same fallback logic) when setting queryContext.sessionId.

Suggested change

	Object sessionId = executionContext.getInputContext().getEnvironmentValue("sessionId");
	if (sessionId != null) {
	queryContext.setSessionId(String.valueOf(sessionId));
	String sessionId = EvaluationLogContextHelper.getSessionId(executionContext.getInputContext());
	if (StringUtils.hasText(sessionId)) {
	queryContext.setSessionId(sessionId);

[Copilot]

experiences.stream().map(...).collect(toList()) is computed multiple times (for logging and for metadata). Consider computing experienceIds once and reusing it to avoid duplicate work and ensure the log/metadata always refer to the same list.