Filter A365 exports to only include genAI spans ADO#37660292

[alexlu4250]

Filter A365 exports to only include genAI spans
ADO#37660292 Task 37660292: [Python] Ensure a365 exports only genAI spans

[github-actions]

⚠️ Deprecation Warning: The deny-licenses option is deprecated for possible removal in the next major release. For more information, see issue 997.

Dependency Review

✅ No vulnerabilities or license issues or OpenSSF Scorecard issues found.

Scanned Files

None

[Copilot]

Pull request overview

This PR updates the Agent365 exporter pipeline so that only spans identified as “genAI” (via gen_ai.operation.name) are eligible for export to the A365 ingest endpoint, with additional normalization for inference operation values to the canonical operation name expected by the service.

Changes:

• Add an allowlist-based filter in exporter utilities to drop spans without an eligible gen_ai.operation.name before partitioning/export.
• Normalize inference operation name enum values (Chat, TextCompletion, GenerateContent) to the canonical "chat" in the exported payload.
• Extend exporter unit tests to validate filtering behavior and operation-name normalization.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.

File	Description
tests/observability/core/test_agent365_exporter.py	Adds tests ensuring non-eligible spans are filtered and inference operation names are normalized in the export payload.
libraries/microsoft-agents-a365-observability-core/microsoft_agents_a365/observability/core/exporters/utils.py	Introduces allowlisted genAI operation-name filtering during identity partitioning and adds related logging.
libraries/microsoft-agents-a365-observability-core/microsoft_agents_a365/observability/core/exporters/agent365_exporter.py	Normalizes inference operation-name enum values to canonical "chat" during span mapping for export.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.

Comments suppressed due to low confidence (1)

libraries/microsoft-agents-a365-observability-core/microsoft_agents_a365/observability/core/exporters/agent365_exporter.py:82

• export() logs "No spans with tenant/agent identity found" when groups is empty, but groups can now be empty because spans were filtered out as non-export-eligible (unknown/missing gen_ai.operation.name), even if tenant/agent identity exists. Consider updating the log message (and corresponding test expectations) to reflect "no export-eligible genAI spans" rather than "no identity" so troubleshooting is accurate.

            if not groups:
                # No spans with identity; treat as success
                logger.info("No spans with tenant/agent identity found; nothing exported.")
                return SpanExportResult.SUCCESS

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated no new comments.

Comments suppressed due to low confidence (1)

libraries/microsoft-agents-a365-observability-core/microsoft_agents_a365/observability/core/exporters/agent365_exporter.py:87

• When filter_and_partition_by_identity() filters out all spans (e.g., missing/unsupported gen_ai.operation.name), groups becomes empty and the exporter logs "No spans with tenant/agent identity found". That message is now misleading because identity may be present but spans were filtered as non-export-eligible. Consider updating the log text (and possibly its level) to reflect "no export-eligible spans" / "all spans filtered out" so operators can diagnose why nothing was exported.

            if not groups:
                # No spans with identity; treat as success
                logger.info("No spans with tenant/agent identity found; nothing exported.")
                return SpanExportResult.SUCCESS

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

[Copilot]

Pull request overview

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

[Copilot]

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.