Move a365 and genai code under microsoft.opentelemetry namespace

pyproject.toml

@@ -49,7 +49,7 @@ dev = [
 ]
 
 [project.entry-points.opentelemetry_instrumentor]
-langchain = "microsoft.genai._langchain._tracer_instrumentor:LangChainInstrumentor"
+langchain = "microsoft.opentelemetry._genai._langchain._tracer_instrumentor:LangChainInstrumentor"
 
 [tool.setuptools]
 package-dir = {"" = "src"}

src/microsoft/opentelemetry/_distro.py

@@ -188,12 +188,12 @@ def _append_a365_components(
         return
 
     try:
-        from microsoft.agents.a365.observability import (
+        from microsoft.opentelemetry.a365 import (
             A365Handlers,
             create_a365_components,
         )
     except ImportError:
-        _logger.warning("A365 export requested but microsoft.agents.a365.observability not available.")
+        _logger.warning("A365 export requested but microsoft.opentelemetry.a365 not available.")
         return
 
     try:

src/microsoft/genai/_langchain/__init__.py → src/microsoft/opentelemetry/_genai/_langchain/__init__.py

@@ -6,6 +6,6 @@
 # package is published to PyPI, conforms to the latest GenAI semantic
 # conventions, and is functionally mature. See PLANNING.md for migration criteria.
 
-from microsoft.genai._langchain._tracer_instrumentor import LangChainInstrumentor
+from microsoft.opentelemetry._genai._langchain._tracer_instrumentor import LangChainInstrumentor
 
 __all__ = ["LangChainInstrumentor"]

src/microsoft/genai/_langchain/_tracer.py → src/microsoft/opentelemetry/_genai/_langchain/_tracer.py

@@ -33,7 +33,7 @@
 from opentelemetry.util.genai.types import Error, LLMInvocation
 from opentelemetry.util.types import AttributeValue
 
-from microsoft.genai._langchain._utils import (
+from microsoft.opentelemetry._genai._langchain._utils import (
     DictWithLock,
     as_utc_nano,
     build_llm_invocation,

src/microsoft/genai/_langchain/_tracer_instrumentor.py → src/microsoft/opentelemetry/_genai/_langchain/_tracer_instrumentor.py

@@ -26,7 +26,7 @@
 from opentelemetry.trace import Span
 from wrapt import wrap_function_wrapper
 
-from microsoft.genai._langchain._tracer import LangChainTracer
+from microsoft.opentelemetry._genai._langchain._tracer import LangChainTracer
 
 logger = logging.getLogger(__name__)
 

src/microsoft/opentelemetry/a365/README.md

@@ -0,0 +1,127 @@
+# A365 Observability (`microsoft.opentelemetry.a365`)
+
+Agent365 observability SDK for OpenTelemetry-based tracing of AI agent operations. Provides scopes, exporters, middleware, and utilities for instrumenting agent invocations, LLM inference calls, tool executions, and output messages.
+
+## Package Structure
+
+### Root
+
+| File | Description |
+|------|-------------|
+| `__init__.py` | Package entry point. Exports `A365Handlers`, `create_a365_components`, and `is_a365_enabled`. |
+| `constants.py` | Distro-level constants: span operation names, OTel semantic convention keys, feature switches, GenAI conventions, tool/user/agent attribute keys, and environment variable names. |
+
+### `core/`
+
+Core tracing primitives — scopes, configuration, data models, and internal utilities.
+
+| File | Description |
+|------|-------------|
+| `__init__.py` | Public API surface. Re-exports 40+ classes/functions from submodules (scopes, models, config, exporters). |
+| `agent_details.py` | `AgentDetails` dataclass — metadata about an AI agent (ID, name, description, blueprint/platform IDs, tenant, version). |
+| `channel.py` | `Channel` dataclass — channel context (name, link) for agent execution. |
+| `config.py` | `TelemetryManager` singleton — thread-safe telemetry configuration. Sets up tracers, span processors, and exporters (Agent365 or Spectra). Provides `configure()`, `get_tracer()`, etc. |
+| `constants.py` | Core-level constants for span operations, OTel conventions, feature switches, and error types. |
+| `execute_tool_scope.py` | `ExecuteToolScope` — tracing scope for AI tool executions. Records tool name, arguments, call ID, type, and endpoint. |
+| `inference_call_details.py` | `InferenceCallDetails` dataclass — LLM call metadata (model, provider, token counts, finish reasons, endpoint). |
+| `inference_operation_type.py` | `InferenceOperationType` enum — Chat, TextCompletion, GenerateContent. |
+| `inference_scope.py` | `InferenceScope` — tracing scope for LLM/AI inference operations. Records input/output messages, model details, token usage, and user info. |
+| `invoke_agent_details.py` | `InvokeAgentScopeDetails` dataclass — configuration for agent invocation tracing (endpoint). |
+| `invoke_agent_scope.py` | `InvokeAgentScope` — tracing scope for agent invocations. Records request/response, caller details (human and agent-to-agent), channel, and endpoint info. |
+| `message_utils.py` | Conversion and serialization helpers for OTel gen-ai message format. Normalizes strings/lists to structured `InputMessages`/`OutputMessages`. |
+| `opentelemetry_scope.py` | `OpenTelemetryScope` — base class for all tracing scopes. Manages span creation, attribute setting, context management, and baggage building. |
+| `request.py` | `Request` dataclass — request details (content, session ID, channel, conversation ID). |
+| `span_details.py` | `SpanDetails` dataclass — span configuration (kind override, parent context, start/end times, links). |
+| `tool_call_details.py` | `ToolCallDetails` dataclass — tool call info (name, arguments, call ID, description, type, endpoint). |
+| `tool_type.py` | `ToolType` enum — Function, Extension, Datastore. |
+| `utils.py` | Utility functions: W3C context extraction, safe JSON serialization, IP validation, baggage manipulation, version retrieval. |
+
+### `core/exporters/`
+
+Span export pipeline — processors and exporters for Agent365 and Spectra backends.
+
+| File | Description |
+|------|-------------|
+| `agent365_exporter.py` | `_Agent365Exporter` — SpanExporter that partitions spans by (tenantId, agentId), builds OTLP-like JSON payloads, and POSTs to Agent365 with Bearer token auth and retry logic. |
+| `agent365_exporter_options.py` | `Agent365ExporterOptions` — configuration for the Agent365 exporter (cluster category, token resolver, endpoint flags, batch settings). |
+| `enriched_span.py` | `EnrichedReadableSpan` — wrapper allowing extra attributes on immutable `ReadableSpan` objects. |
+| `enriching_span_processor.py` | Span enrichment support with registration for platform instrumentors (LangChain, Semantic Kernel, OpenAI Agents). `_EnrichingBatchSpanProcessor` applies enrichers before batching. |
+| `span_processor.py` | `A365SpanProcessor` — propagates OpenTelemetry baggage entries onto spans as attributes, with special handling for invoke_agent spans. |
+| `spectra_exporter_options.py` | `SpectraExporterOptions` — configuration for OTLP export to a Spectra Collector sidecar (gRPC or HTTP, tuned for Kubernetes). |
+| `utils.py` | Exporter utilities: hex encoding for trace/span IDs, span size truncation, span partitioning, environment variable handling, payload building helpers. |
+
+### `core/middleware/`
+
+| File | Description |
+|------|-------------|
+| `baggage_builder.py` | `BaggageBuilder` — fluent API for setting per-request baggage values (tenant ID, agent ID, caller/user details, session/conversation IDs, channel, endpoints). Provides context manager for baggage scope. |
+
+### `core/models/`
+
+Data models for messages, agents, callers, and responses.
+
+| File | Description |
+|------|-------------|
+| `agent_type.py` | `AgentType` enum — EntraEmbodied, EntraNonEmbodied, MicrosoftCopilot, DeclarativeAgent, Foundry. |
+| `caller_details.py` | `CallerDetails` dataclass — groups caller identity for agent-to-agent (A2A) scenarios (human user and calling agent). |
+| `messages.py` | OTel gen-ai message types: enums (`MessageRole`, `FinishReason`, `Modality`), message parts (`TextPart`, `BlobPart`, `FilePart`, `UriPart`), `ChatMessage`, `OutputMessage`, and wrapper types. |
+| `operation_source.py` | `OperationSource` enum — SDK, Gateway, MCPServer. |
+| `response.py` | `Response` dataclass — output messages (strings, `OutputMessages`, or tool result dicts). |
+| `service_endpoint.py` | `ServiceEndpoint` dataclass — hostname and optional port. |
+| `user_details.py` | `UserDetails` dataclass — human user info (ID, email, name, client IP). |
+
+### `core/spans_scopes/`
+
+| File | Description |
+|------|-------------|
+| `output_scope.py` | `OutputScope` — tracing scope for output messages. Records output with agent/user details and conversation ID. |
+
+### `core/trace_processor/`
+
+| File | Description |
+|------|-------------|
+| `span_processor.py` | `SpanProcessor` — copies OpenTelemetry baggage entries to span attributes with special handling for invoke_agent spans. |
+| `util.py` | Attribute lists for baggage-to-span propagation: `COMMON_ATTRIBUTES` and `INVOKE_AGENT_ATTRIBUTES`. |
+
+### `hosting/`
+
+Hosting-layer middleware for Bot Framework integration.
+
+| File | Description |
+|------|-------------|
+| `__init__.py` | Exports `BaggageMiddleware`, `OutputLoggingMiddleware`, `ObservabilityHostingManager`, `ObservabilityHostingOptions`. |
+
+### `hosting/middleware/`
+
+| File | Description |
+|------|-------------|
+| `baggage_middleware.py` | `BaggageMiddleware` — propagates OpenTelemetry baggage context from `TurnContext`. Skips async ContinueConversation events. |
+| `observability_hosting_manager.py` | `ObservabilityHostingManager` singleton — configures hosting-layer middleware. `ObservabilityHostingOptions` dataclass for enabling/disabling middleware. |
+| `output_logging_middleware.py` | `OutputLoggingMiddleware` — creates `OutputScope` spans for outgoing messages. Links to parent span via traceparent in turn_state. |
+
+### `hosting/scope_helpers/`
+
+| File | Description |
+|------|-------------|
+| `populate_baggage.py` | Extracts baggage pairs from `TurnContext` activity and sets them on a `BaggageBuilder`. |
+| `populate_invoke_agent_scope.py` | Extracts `TurnContext` activity data and records as `InvokeAgentScope` attributes and input messages. |
+| `utils.py` | Pair extraction helpers for `TurnContext` activity: caller pairs, target agent pairs, tenant ID, channel, conversation. |
+
+### `hosting/token_cache_helpers/`
+
+| File | Description |
+|------|-------------|
+| `agent_token_cache.py` | `AgenticTokenCache` — thread-safe cache for observability tokens per (agentId, tenantId). `AgenticTokenStruct` wraps Authorization and TurnContext for token generation. |
+
+### `runtime/`
+
+Runtime utilities for environment discovery, error handling, and token operations.
+
+| File | Description |
+|------|-------------|
+| `environment_utils.py` | Environment utilities: observability authentication scopes (with override support), dev environment detection. |
+| `operation_error.py` | `OperationError` — wraps an exception with message property. |
+| `operation_result.py` | `OperationResult` — represents operation success/failure with optional error list. Singleton success instance. |
+| `power_platform_api_discovery.py` | `PowerPlatformApiDiscovery` — discovers Power Platform API endpoints per cluster category. `ClusterCategory` type for valid cluster strings. |
+| `utility.py` | `Utility` — static methods for JWT token decoding (diagnostics only), extracting app IDs and agent blueprint IDs from tokens. |
+| `version_utils.py` | Deprecated versioning utilities. `build_version()` deprecated in favor of setuptools-git-versioning. |

src/microsoft/agents/a365/observability/__init__.py → src/microsoft/opentelemetry/a365/__init__.py

@@ -6,7 +6,7 @@
 
 """Agent365 exporter integration for the Microsoft OpenTelemetry Distro."""
 
-from microsoft.agents.a365.observability.core.exporters.utils import (  # noqa: F401
+from microsoft.opentelemetry.a365.core.exporters.utils import (  # noqa: F401
     A365Handlers,
     create_a365_components,
     is_a365_enabled,

src/microsoft/agents/a365/observability/constants.py → src/microsoft/opentelemetry/a365/constants.py

@@ -45,8 +45,8 @@
 GEN_AI_AGENT_NAME_KEY = "gen_ai.agent.name"
 GEN_AI_AGENT_DESCRIPTION_KEY = "gen_ai.agent.description"
 GEN_AI_AGENT_VERSION_KEY = "gen_ai.agent.version"
-GEN_AI_AGENT_PLATFORM_ID_KEY = "microsoft.a365.agent.platform.id"
-GEN_AI_AGENT_THOUGHT_PROCESS_KEY = "microsoft.a365.agent.thought.process"
+GEN_AI_AGENT_PLATFORM_ID_KEY = "microsoft.opentelemetry.a365.agent.platform.id"
+GEN_AI_AGENT_THOUGHT_PROCESS_KEY = "microsoft.opentelemetry.a365.agent.thought.process"
 GEN_AI_CONVERSATION_ID_KEY = "gen_ai.conversation.id"
 GEN_AI_CONVERSATION_ITEM_LINK_KEY = "microsoft.conversation.item.link"
 GEN_AI_TOKEN_TYPE_KEY = "gen_ai.token.type"
@@ -73,13 +73,13 @@
 GEN_AI_CALLER_CLIENT_IP_KEY = "client.address"
 
 # --- Agent-to-Agent caller ---
-GEN_AI_CALLER_AGENT_USER_ID_KEY = "microsoft.a365.caller.agent.user.id"
-GEN_AI_CALLER_AGENT_EMAIL_KEY = "microsoft.a365.caller.agent.user.email"
-GEN_AI_CALLER_AGENT_NAME_KEY = "microsoft.a365.caller.agent.name"
-GEN_AI_CALLER_AGENT_ID_KEY = "microsoft.a365.caller.agent.id"
-GEN_AI_CALLER_AGENT_APPLICATION_ID_KEY = "microsoft.a365.caller.agent.blueprint.id"
-GEN_AI_CALLER_AGENT_PLATFORM_ID_KEY = "microsoft.a365.caller.agent.platform.id"
-GEN_AI_CALLER_AGENT_VERSION_KEY = "microsoft.a365.caller.agent.version"
+GEN_AI_CALLER_AGENT_USER_ID_KEY = "microsoft.opentelemetry.a365.caller.agent.user.id"
+GEN_AI_CALLER_AGENT_EMAIL_KEY = "microsoft.opentelemetry.a365.caller.agent.user.email"
+GEN_AI_CALLER_AGENT_NAME_KEY = "microsoft.opentelemetry.a365.caller.agent.name"
+GEN_AI_CALLER_AGENT_ID_KEY = "microsoft.opentelemetry.a365.caller.agent.id"
+GEN_AI_CALLER_AGENT_APPLICATION_ID_KEY = "microsoft.opentelemetry.a365.caller.agent.blueprint.id"
+GEN_AI_CALLER_AGENT_PLATFORM_ID_KEY = "microsoft.opentelemetry.a365.caller.agent.platform.id"
+GEN_AI_CALLER_AGENT_VERSION_KEY = "microsoft.opentelemetry.a365.caller.agent.version"
 
 # --- Agent-specific dimensions ---
 AGENT_ID_KEY = "gen_ai.agent.id"
@@ -89,7 +89,7 @@
 TENANT_ID_KEY = "microsoft.tenant.id"
 GEN_AI_AGENT_AUID_KEY = "microsoft.agent.user.id"
 GEN_AI_AGENT_EMAIL_KEY = "microsoft.agent.user.email"
-GEN_AI_AGENT_BLUEPRINT_ID_KEY = "microsoft.a365.agent.blueprint.id"
+GEN_AI_AGENT_BLUEPRINT_ID_KEY = "microsoft.opentelemetry.a365.agent.blueprint.id"
 SESSION_ID_KEY = "microsoft.session.id"
 SESSION_DESCRIPTION_KEY = "microsoft.session.description"