chore!: remove telemetry API usage (#3815)

# What does this PR do?

remove telemetry as a providable API from the codebase. This includes
removing it from generated distributions but also the provider registry,
the router, etc

since `setup_logger` is tied pretty strictly to `Api.telemetry` being in
impls we still need an "instantiated provider" in our implementations.
However it should not be auto-routed or provided. So in
validate_and_prepare_providers (called from resolve_impls) I made it so
that if run_config.telemetry.enabled, we set up the meta-reference
"provider" internally to be used so that log_event will work when
called.

This is the neatest way I think we can remove telemetry from the
provider configs but also not need to rip apart the whole "telemetry is
a provider" logic just yet, but we can do it internally later without
disrupting users.

so telemetry is removed from the registry such that if a user puts
`telemetry:` as an API in their build/run config it will err out, but
can still be used by us internally as we go through this transition.


relates to #3806

Signed-off-by: Charlie Doern <cdoern@redhat.com>

docs/docs/distributions/self_hosted_distro/meta-reference-gpu.md

@@ -21,7 +21,6 @@ The `llamastack/distribution-meta-reference-gpu` distribution consists of the fo
 | inference | `inline::meta-reference` |
 | safety | `inline::llama-guard` |
 | scoring | `inline::basic`, `inline::llm-as-judge`, `inline::braintrust` |
-| telemetry | `inline::meta-reference` |
 | tool_runtime | `remote::brave-search`, `remote::tavily-search`, `inline::rag-runtime`, `remote::model-context-protocol` |
 | vector_io | `inline::faiss`, `remote::chromadb`, `remote::pgvector` |
 

docs/docs/distributions/self_hosted_distro/nvidia.md

@@ -16,7 +16,6 @@ The `llamastack/distribution-nvidia` distribution consists of the following prov
 | post_training | `remote::nvidia` |
 | safety | `remote::nvidia` |
 | scoring | `inline::basic` |
-| telemetry | `inline::meta-reference` |
 | tool_runtime | `inline::rag-runtime` |
 | vector_io | `inline::faiss` |
 

llama_stack/core/configure.py

@@ -64,7 +64,7 @@ def configure_api_providers(config: StackRunConfig, build_spec: DistributionSpec
     if config.apis:
         apis_to_serve = config.apis
     else:
-        apis_to_serve = [a.value for a in Api if a not in (Api.telemetry, Api.inspect, Api.providers)]
+        apis_to_serve = [a.value for a in Api if a not in (Api.inspect, Api.providers)]
 
     for api_str in apis_to_serve:
         api = Api(api_str)

llama_stack/core/datatypes.py

@@ -176,6 +176,10 @@ class DistributionSpec(BaseModel):
     )
 
 
+class TelemetryConfig(BaseModel):
+    enabled: bool = Field(default=False, description="enable or disable telemetry")
+
+
 class LoggingConfig(BaseModel):
     category_levels: dict[str, str] = Field(
         default_factory=dict,
@@ -493,6 +497,8 @@ If not specified, a default SQLite store will be used.""",
 
     logging: LoggingConfig | None = Field(default=None, description="Configuration for Llama Stack Logging")
 
+    telemetry: TelemetryConfig = Field(default_factory=TelemetryConfig, description="Configuration for telemetry")
+
     server: ServerConfig = Field(
         default_factory=ServerConfig,
         description="Configuration for the HTTP(S) server",

llama_stack/core/distribution.py

@@ -25,7 +25,7 @@ from llama_stack.providers.datatypes import (
 logger = get_logger(name=__name__, category="core")
 
 
-INTERNAL_APIS = {Api.inspect, Api.providers, Api.prompts, Api.conversations}
+INTERNAL_APIS = {Api.inspect, Api.providers, Api.prompts, Api.conversations, Api.telemetry}
 
 
 def stack_apis() -> list[Api]:

llama_stack/core/library_client.py

@@ -48,12 +48,7 @@ from llama_stack.core.utils.config import redact_sensitive_fields
 from llama_stack.core.utils.context import preserve_contexts_async_generator
 from llama_stack.core.utils.exec import in_notebook
 from llama_stack.log import get_logger
-from llama_stack.providers.utils.telemetry.tracing import (
-    CURRENT_TRACE_CONTEXT,
-    end_trace,
-    setup_logger,
-    start_trace,
-)
+from llama_stack.providers.utils.telemetry.tracing import CURRENT_TRACE_CONTEXT, end_trace, setup_logger, start_trace
 from llama_stack.strong_typing.inspection import is_unwrapped_body_param
 
 logger = get_logger(name=__name__, category="core")

llama_stack/core/resolver.py

@@ -26,7 +26,6 @@ from llama_stack.apis.safety import Safety
 from llama_stack.apis.scoring import Scoring
 from llama_stack.apis.scoring_functions import ScoringFunctions
 from llama_stack.apis.shields import Shields
-from llama_stack.apis.telemetry import Telemetry
 from llama_stack.apis.tools import ToolGroups, ToolRuntime
 from llama_stack.apis.vector_io import VectorIO
 from llama_stack.apis.version import LLAMA_STACK_API_V1ALPHA
@@ -47,6 +46,7 @@ from llama_stack.providers.datatypes import (
     Api,
     BenchmarksProtocolPrivate,
     DatasetsProtocolPrivate,
+    InlineProviderSpec,
     ModelsProtocolPrivate,
     ProviderSpec,
     RemoteProviderConfig,
@@ -82,7 +82,6 @@ def api_protocol_map(external_apis: dict[Api, ExternalApiSpec] | None = None) ->
         Api.models: Models,
         Api.safety: Safety,
         Api.shields: Shields,
-        Api.telemetry: Telemetry,
         Api.datasetio: DatasetIO,
         Api.datasets: Datasets,
         Api.scoring: Scoring,
@@ -204,9 +203,7 @@ def specs_for_autorouted_apis(apis_to_serve: list[str] | set[str]) -> dict[str,
                     module="llama_stack.core.routers",
                     routing_table_api=info.routing_table_api,
                     api_dependencies=[info.routing_table_api],
-                    # Add telemetry as an optional dependency to all auto-routed providers
-                    optional_api_dependencies=[Api.telemetry],
-                    deps__=([info.routing_table_api.value, Api.telemetry.value]),
+                    deps__=([info.routing_table_api.value]),
                 ),
             )
         }
@@ -239,6 +236,24 @@ def validate_and_prepare_providers(
         key = api_str if api not in router_apis else f"inner-{api_str}"
         providers_with_specs[key] = specs
 
+    # TODO: remove this logic, telemetry should not have providers.
+    # if telemetry has been enabled in the config initialize our internal impl
+    # telemetry is not an external API so it SHOULD NOT be auto-routed.
+    if run_config.telemetry.enabled:
+        specs = {}
+        p = InlineProviderSpec(
+            api=Api.telemetry,
+            provider_type="inline::meta-reference",
+            pip_packages=[],
+            optional_api_dependencies=[Api.datasetio],
+            module="llama_stack.providers.inline.telemetry.meta_reference",
+            config_class="llama_stack.providers.inline.telemetry.meta_reference.config.TelemetryConfig",
+            description="Meta's reference implementation of telemetry and observability using OpenTelemetry.",
+        )
+        spec = ProviderWithSpec(spec=p, provider_type="inline::meta-reference", provider_id="meta-reference")
+        specs["meta-reference"] = spec
+        providers_with_specs["telemetry"] = specs
+
     return providers_with_specs
 
 
@@ -389,6 +404,8 @@ async def instantiate_provider(
         args = [config, deps]
         if "policy" in inspect.signature(getattr(module, method)).parameters:
             args.append(policy)
+        if "telemetry_enabled" in inspect.signature(getattr(module, method)).parameters and run_config.telemetry:
+            args.append(run_config.telemetry.enabled)
 
     fn = getattr(module, method)
     impl = await fn(*args)

llama_stack/core/routers/__init__.py

@@ -63,16 +63,17 @@ async def get_auto_router_impl(
         "eval": EvalRouter,
         "tool_runtime": ToolRuntimeRouter,
     }
-    api_to_deps = {
-        "inference": {"telemetry": Api.telemetry},
-    }
     if api.value not in api_to_routers:
         raise ValueError(f"API {api.value} not found in router map")
 
     api_to_dep_impl = {}
-    for dep_name, dep_api in api_to_deps.get(api.value, {}).items():
-        if dep_api in deps:
-            api_to_dep_impl[dep_name] = deps[dep_api]
+    if run_config.telemetry.enabled:
+        api_to_deps = {
+            "inference": {"telemetry": Api.telemetry},
+        }
+        for dep_name, dep_api in api_to_deps.get(api.value, {}).items():
+            if dep_api in deps:
+                api_to_dep_impl[dep_name] = deps[dep_api]
 
     # TODO: move pass configs to routers instead
     if api == Api.inference and run_config.inference_store:

llama_stack/core/server/server.py

@@ -445,7 +445,7 @@ def create_app() -> StackApp:
         if cors_config:
             app.add_middleware(CORSMiddleware, **cors_config.model_dump())
 
-    if Api.telemetry in impls:
+    if config.telemetry.enabled:
         setup_logger(impls[Api.telemetry])
     else:
         setup_logger(TelemetryAdapter(TelemetryConfig(), {}))

llama_stack/distributions/ci-tests/build.yaml

@@ -32,8 +32,6 @@ distribution_spec:
     - provider_type: inline::code-scanner
     agents:
     - provider_type: inline::meta-reference
-    telemetry:
-    - provider_type: inline::meta-reference
     post_training:
     - provider_type: inline::torchtune-cpu
     eval:

llama_stack/distributions/ci-tests/run.yaml

@@ -10,7 +10,6 @@ apis:
 - post_training
 - safety
 - scoring
-- telemetry
 - tool_runtime
 - vector_io
 providers:
@@ -154,13 +153,6 @@ providers:
       responses_store:
         type: sqlite
         db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/ci-tests}/responses_store.db
-  telemetry:
-  - provider_id: meta-reference
-    provider_type: inline::meta-reference
-    config:
-      service_name: "${env.OTEL_SERVICE_NAME:=\u200B}"
-      sinks: ${env.TELEMETRY_SINKS:=}
-      otel_exporter_otlp_endpoint: ${env.OTEL_EXPORTER_OTLP_ENDPOINT:=}
   post_training:
   - provider_id: torchtune-cpu
     provider_type: inline::torchtune-cpu

llama_stack/distributions/dell/build.yaml

@@ -14,8 +14,6 @@ distribution_spec:
     - provider_type: inline::llama-guard
     agents:
     - provider_type: inline::meta-reference
-    telemetry:
-    - provider_type: inline::meta-reference
     eval:
     - provider_type: inline::meta-reference
     datasetio:

llama_stack/distributions/dell/dell.py

@@ -32,7 +32,6 @@ def get_distribution_template() -> DistributionTemplate:
         ],
         "safety": [BuildProvider(provider_type="inline::llama-guard")],
         "agents": [BuildProvider(provider_type="inline::meta-reference")],
-        "telemetry": [BuildProvider(provider_type="inline::meta-reference")],
         "eval": [BuildProvider(provider_type="inline::meta-reference")],
         "datasetio": [
             BuildProvider(provider_type="remote::huggingface"),

llama_stack/distributions/dell/run-with-safety.yaml

@@ -7,7 +7,6 @@ apis:
 - inference
 - safety
 - scoring
-- telemetry
 - tool_runtime
 - vector_io
 providers:
@@ -45,13 +44,6 @@ providers:
       responses_store:
         type: sqlite
         db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/dell}/responses_store.db
-  telemetry:
-  - provider_id: meta-reference
-    provider_type: inline::meta-reference
-    config:
-      service_name: "${env.OTEL_SERVICE_NAME:=\u200B}"
-      sinks: ${env.TELEMETRY_SINKS:=}
-      otel_exporter_otlp_endpoint: ${env.OTEL_EXPORTER_OTLP_ENDPOINT:=}
   eval:
   - provider_id: meta-reference
     provider_type: inline::meta-reference

llama_stack/distributions/dell/run.yaml

@@ -7,7 +7,6 @@ apis:
 - inference
 - safety
 - scoring
-- telemetry
 - tool_runtime
 - vector_io
 providers:
@@ -41,13 +40,6 @@ providers:
       responses_store:
         type: sqlite
         db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/dell}/responses_store.db
-  telemetry:
-  - provider_id: meta-reference
-    provider_type: inline::meta-reference
-    config:
-      service_name: "${env.OTEL_SERVICE_NAME:=\u200B}"
-      sinks: ${env.TELEMETRY_SINKS:=}
-      otel_exporter_otlp_endpoint: ${env.OTEL_EXPORTER_OTLP_ENDPOINT:=}
   eval:
   - provider_id: meta-reference
     provider_type: inline::meta-reference

llama_stack/distributions/meta-reference-gpu/build.yaml

@@ -12,8 +12,6 @@ distribution_spec:
     - provider_type: inline::llama-guard
     agents:
     - provider_type: inline::meta-reference
-    telemetry:
-    - provider_type: inline::meta-reference
     eval:
     - provider_type: inline::meta-reference
     datasetio:

llama_stack/distributions/meta-reference-gpu/meta_reference.py

@@ -34,7 +34,6 @@ def get_distribution_template() -> DistributionTemplate:
         ],
         "safety": [BuildProvider(provider_type="inline::llama-guard")],
         "agents": [BuildProvider(provider_type="inline::meta-reference")],
-        "telemetry": [BuildProvider(provider_type="inline::meta-reference")],
         "eval": [BuildProvider(provider_type="inline::meta-reference")],
         "datasetio": [
             BuildProvider(provider_type="remote::huggingface"),

llama_stack/distributions/meta-reference-gpu/run-with-safety.yaml

@@ -7,7 +7,6 @@ apis:
 - inference
 - safety
 - scoring
-- telemetry
 - tool_runtime
 - vector_io
 providers:
@@ -56,13 +55,6 @@ providers:
       responses_store:
         type: sqlite
         db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/meta-reference-gpu}/responses_store.db
-  telemetry:
-  - provider_id: meta-reference
-    provider_type: inline::meta-reference
-    config:
-      service_name: "${env.OTEL_SERVICE_NAME:=\u200B}"
-      sinks: ${env.TELEMETRY_SINKS:=}
-      otel_exporter_otlp_endpoint: ${env.OTEL_EXPORTER_OTLP_ENDPOINT:=}
   eval:
   - provider_id: meta-reference
     provider_type: inline::meta-reference

llama_stack/distributions/meta-reference-gpu/run.yaml

@@ -7,7 +7,6 @@ apis:
 - inference
 - safety
 - scoring
-- telemetry
 - tool_runtime
 - vector_io
 providers:
@@ -46,13 +45,6 @@ providers:
       responses_store:
         type: sqlite
         db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/meta-reference-gpu}/responses_store.db
-  telemetry:
-  - provider_id: meta-reference
-    provider_type: inline::meta-reference
-    config:
-      service_name: "${env.OTEL_SERVICE_NAME:=\u200B}"
-      sinks: ${env.TELEMETRY_SINKS:=}
-      otel_exporter_otlp_endpoint: ${env.OTEL_EXPORTER_OTLP_ENDPOINT:=}
   eval:
   - provider_id: meta-reference
     provider_type: inline::meta-reference

llama_stack/distributions/nvidia/build.yaml

@@ -10,8 +10,6 @@ distribution_spec:
     - provider_type: remote::nvidia
     agents:
     - provider_type: inline::meta-reference
-    telemetry:
-    - provider_type: inline::meta-reference
     eval:
     - provider_type: remote::nvidia
     post_training:

llama_stack/distributions/nvidia/nvidia.py

@@ -21,7 +21,6 @@ def get_distribution_template(name: str = "nvidia") -> DistributionTemplate:
         "vector_io": [BuildProvider(provider_type="inline::faiss")],
         "safety": [BuildProvider(provider_type="remote::nvidia")],
         "agents": [BuildProvider(provider_type="inline::meta-reference")],
-        "telemetry": [BuildProvider(provider_type="inline::meta-reference")],
         "eval": [BuildProvider(provider_type="remote::nvidia")],
         "post_training": [BuildProvider(provider_type="remote::nvidia")],
         "datasetio": [

llama_stack/distributions/nvidia/run-with-safety.yaml

@@ -9,7 +9,6 @@ apis:
 - post_training
 - safety
 - scoring
-- telemetry
 - tool_runtime
 - vector_io
 providers:
@@ -48,13 +47,6 @@ providers:
       responses_store:
         type: sqlite
         db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/nvidia}/responses_store.db
-  telemetry:
-  - provider_id: meta-reference
-    provider_type: inline::meta-reference
-    config:
-      service_name: "${env.OTEL_SERVICE_NAME:=\u200B}"
-      sinks: ${env.TELEMETRY_SINKS:=}
-      otel_exporter_otlp_endpoint: ${env.OTEL_EXPORTER_OTLP_ENDPOINT:=}
   eval:
   - provider_id: nvidia
     provider_type: remote::nvidia

llama_stack/distributions/nvidia/run.yaml

@@ -9,7 +9,6 @@ apis:
 - post_training
 - safety
 - scoring
-- telemetry
 - tool_runtime
 - vector_io
 providers:
@@ -43,13 +42,6 @@ providers:
       responses_store:
         type: sqlite
         db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/nvidia}/responses_store.db
-  telemetry:
-  - provider_id: meta-reference
-    provider_type: inline::meta-reference
-    config:
-      service_name: "${env.OTEL_SERVICE_NAME:=\u200B}"
-      sinks: ${env.TELEMETRY_SINKS:=}
-      otel_exporter_otlp_endpoint: ${env.OTEL_EXPORTER_OTLP_ENDPOINT:=}
   eval:
   - provider_id: nvidia
     provider_type: remote::nvidia

llama_stack/distributions/open-benchmark/build.yaml

@@ -16,8 +16,6 @@ distribution_spec:
     - provider_type: inline::llama-guard
     agents:
     - provider_type: inline::meta-reference
-    telemetry:
-    - provider_type: inline::meta-reference
     eval:
     - provider_type: inline::meta-reference
     datasetio:

llama_stack/distributions/open-benchmark/open_benchmark.py

@@ -105,7 +105,6 @@ def get_distribution_template() -> DistributionTemplate:
         ],
         "safety": [BuildProvider(provider_type="inline::llama-guard")],
         "agents": [BuildProvider(provider_type="inline::meta-reference")],
-        "telemetry": [BuildProvider(provider_type="inline::meta-reference")],
         "eval": [BuildProvider(provider_type="inline::meta-reference")],
         "datasetio": [
             BuildProvider(provider_type="remote::huggingface"),

llama_stack/distributions/open-benchmark/run.yaml

@@ -7,7 +7,6 @@ apis:
 - inference
 - safety
 - scoring
-- telemetry
 - tool_runtime
 - vector_io
 providers:
@@ -76,13 +75,6 @@ providers:
       responses_store:
         type: sqlite
         db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/open-benchmark}/responses_store.db
-  telemetry:
-  - provider_id: meta-reference
-    provider_type: inline::meta-reference
-    config:
-      service_name: "${env.OTEL_SERVICE_NAME:=\u200B}"
-      sinks: ${env.TELEMETRY_SINKS:=}
-      otel_exporter_otlp_endpoint: ${env.OTEL_EXPORTER_OTLP_ENDPOINT:=}
   eval:
   - provider_id: meta-reference
     provider_type: inline::meta-reference

llama_stack/distributions/postgres-demo/build.yaml

@@ -11,8 +11,6 @@ distribution_spec:
     - provider_type: inline::llama-guard
     agents:
     - provider_type: inline::meta-reference
-    telemetry:
-    - provider_type: inline::meta-reference
     tool_runtime:
     - provider_type: remote::brave-search
     - provider_type: remote::tavily-search

llama_stack/distributions/postgres-demo/postgres_demo.py

@@ -42,7 +42,6 @@ def get_distribution_template() -> DistributionTemplate:
         "vector_io": [BuildProvider(provider_type="remote::chromadb")],
         "safety": [BuildProvider(provider_type="inline::llama-guard")],
         "agents": [BuildProvider(provider_type="inline::meta-reference")],
-        "telemetry": [BuildProvider(provider_type="inline::meta-reference")],
         "tool_runtime": [
             BuildProvider(provider_type="remote::brave-search"),
             BuildProvider(provider_type="remote::tavily-search"),
@@ -116,17 +115,6 @@ def get_distribution_template() -> DistributionTemplate:
                             ),
                         )
                     ],
-                    "telemetry": [
-                        Provider(
-                            provider_id="meta-reference",
-                            provider_type="inline::meta-reference",
-                            config=dict(
-                                service_name="${env.OTEL_SERVICE_NAME:=\u200b}",
-                                sinks="${env.TELEMETRY_SINKS:=console,otel_trace}",
-                                otel_exporter_otlp_endpoint="${env.OTEL_EXPORTER_OTLP_ENDPOINT:=http://localhost:4318/v1/traces}",
-                            ),
-                        )
-                    ],
                 },
                 default_models=default_models + [embedding_model],
                 default_tool_groups=default_tool_groups,

llama_stack/distributions/postgres-demo/run.yaml

@@ -4,7 +4,6 @@ apis:
 - agents
 - inference
 - safety
-- telemetry
 - tool_runtime
 - vector_io
 providers:
@@ -49,13 +48,6 @@ providers:
         db: ${env.POSTGRES_DB:=llamastack}
         user: ${env.POSTGRES_USER:=llamastack}
         password: ${env.POSTGRES_PASSWORD:=llamastack}
-  telemetry:
-  - provider_id: meta-reference
-    provider_type: inline::meta-reference
-    config:
-      service_name: "${env.OTEL_SERVICE_NAME:=\u200B}"
-      sinks: ${env.TELEMETRY_SINKS:=console,otel_trace}
-      otel_exporter_otlp_endpoint: ${env.OTEL_EXPORTER_OTLP_ENDPOINT:=http://localhost:4318/v1/traces}
   tool_runtime:
   - provider_id: brave-search
     provider_type: remote::brave-search

llama_stack/distributions/starter-gpu/build.yaml

@@ -33,8 +33,6 @@ distribution_spec:
     - provider_type: inline::code-scanner
     agents:
     - provider_type: inline::meta-reference
-    telemetry:
-    - provider_type: inline::meta-reference
     post_training:
     - provider_type: inline::huggingface-gpu
     eval:

llama_stack/distributions/starter-gpu/run.yaml

@@ -10,7 +10,6 @@ apis:
 - post_training
 - safety
 - scoring
-- telemetry
 - tool_runtime
 - vector_io
 providers:
@@ -154,13 +153,6 @@ providers:
       responses_store:
         type: sqlite
         db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/starter-gpu}/responses_store.db
-  telemetry:
-  - provider_id: meta-reference
-    provider_type: inline::meta-reference
-    config:
-      service_name: "${env.OTEL_SERVICE_NAME:=\u200B}"
-      sinks: ${env.TELEMETRY_SINKS:=}
-      otel_exporter_otlp_endpoint: ${env.OTEL_EXPORTER_OTLP_ENDPOINT:=}
   post_training:
   - provider_id: huggingface-gpu
     provider_type: inline::huggingface-gpu

llama_stack/distributions/starter/build.yaml

@@ -33,8 +33,6 @@ distribution_spec:
     - provider_type: inline::code-scanner
     agents:
     - provider_type: inline::meta-reference
-    telemetry:
-    - provider_type: inline::meta-reference
     post_training:
     - provider_type: inline::torchtune-cpu
     eval:

llama_stack/distributions/starter/run.yaml

@@ -10,7 +10,6 @@ apis:
 - post_training
 - safety
 - scoring
-- telemetry
 - tool_runtime
 - vector_io
 providers:
@@ -154,13 +153,6 @@ providers:
       responses_store:
         type: sqlite
         db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/starter}/responses_store.db
-  telemetry:
-  - provider_id: meta-reference
-    provider_type: inline::meta-reference
-    config:
-      service_name: "${env.OTEL_SERVICE_NAME:=\u200B}"
-      sinks: ${env.TELEMETRY_SINKS:=}
-      otel_exporter_otlp_endpoint: ${env.OTEL_EXPORTER_OTLP_ENDPOINT:=}
   post_training:
   - provider_id: torchtune-cpu
     provider_type: inline::torchtune-cpu

llama_stack/distributions/starter/starter.py

@@ -120,7 +120,6 @@ def get_distribution_template(name: str = "starter") -> DistributionTemplate:
             BuildProvider(provider_type="inline::code-scanner"),
         ],
         "agents": [BuildProvider(provider_type="inline::meta-reference")],
-        "telemetry": [BuildProvider(provider_type="inline::meta-reference")],
         "post_training": [BuildProvider(provider_type="inline::torchtune-cpu")],
         "eval": [BuildProvider(provider_type="inline::meta-reference")],
         "datasetio": [

llama_stack/distributions/watsonx/build.yaml

@@ -11,8 +11,6 @@ distribution_spec:
     - provider_type: inline::llama-guard
     agents:
     - provider_type: inline::meta-reference
-    telemetry:
-    - provider_type: inline::meta-reference
     eval:
     - provider_type: inline::meta-reference
     datasetio:

llama_stack/distributions/watsonx/run.yaml

@@ -8,7 +8,6 @@ apis:
 - inference
 - safety
 - scoring
-- telemetry
 - tool_runtime
 - vector_io
 providers:
@@ -41,13 +40,6 @@ providers:
       responses_store:
         type: sqlite
         db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/watsonx}/responses_store.db
-  telemetry:
-  - provider_id: meta-reference
-    provider_type: inline::meta-reference
-    config:
-      service_name: "${env.OTEL_SERVICE_NAME:=\u200B}"
-      sinks: ${env.TELEMETRY_SINKS:=}
-      otel_exporter_otlp_endpoint: ${env.OTEL_EXPORTER_OTLP_ENDPOINT:=}
   eval:
   - provider_id: meta-reference
     provider_type: inline::meta-reference

llama_stack/distributions/watsonx/watsonx.py

@@ -20,7 +20,6 @@ def get_distribution_template(name: str = "watsonx") -> DistributionTemplate:
         "vector_io": [BuildProvider(provider_type="inline::faiss")],
         "safety": [BuildProvider(provider_type="inline::llama-guard")],
         "agents": [BuildProvider(provider_type="inline::meta-reference")],
-        "telemetry": [BuildProvider(provider_type="inline::meta-reference")],
         "eval": [BuildProvider(provider_type="inline::meta-reference")],
         "datasetio": [
             BuildProvider(provider_type="remote::huggingface"),

llama_stack/providers/inline/agents/meta_reference/__init__.py

@@ -11,7 +11,12 @@ from llama_stack.core.datatypes import AccessRule, Api
 from .config import MetaReferenceAgentsImplConfig
 
 
-async def get_provider_impl(config: MetaReferenceAgentsImplConfig, deps: dict[Api, Any], policy: list[AccessRule]):
+async def get_provider_impl(
+    config: MetaReferenceAgentsImplConfig,
+    deps: dict[Api, Any],
+    policy: list[AccessRule],
+    telemetry_enabled: bool = False,
+):
     from .agents import MetaReferenceAgentsImpl
 
     impl = MetaReferenceAgentsImpl(
@@ -23,7 +28,7 @@ async def get_provider_impl(config: MetaReferenceAgentsImplConfig, deps: dict[Ap
         deps[Api.tool_groups],
         deps[Api.conversations],
         policy,
-        Api.telemetry in deps,
+        telemetry_enabled,
     )
     await impl.initialize()
     return impl

llama_stack/providers/registry/agents.py

@@ -36,9 +36,6 @@ def available_providers() -> list[ProviderSpec]:
                 Api.tool_groups,
                 Api.conversations,
             ],
-            optional_api_dependencies=[
-                Api.telemetry,
-            ],
             description="Meta's reference implementation of an agent system that can use tools, access vector databases, and perform complex reasoning tasks.",
         ),
     ]

llama_stack/providers/registry/telemetry.py

@@ -1,29 +0,0 @@
-# Copyright (c) Meta Platforms, Inc. and affiliates.
-# All rights reserved.
-#
-# This source code is licensed under the terms described in the LICENSE file in
-# the root directory of this source tree.
-
-
-from llama_stack.providers.datatypes import (
-    Api,
-    InlineProviderSpec,
-    ProviderSpec,
-)
-
-
-def available_providers() -> list[ProviderSpec]:
-    return [
-        InlineProviderSpec(
-            api=Api.telemetry,
-            provider_type="inline::meta-reference",
-            pip_packages=[
-                "opentelemetry-sdk",
-                "opentelemetry-exporter-otlp-proto-http",
-            ],
-            optional_api_dependencies=[Api.datasetio],
-            module="llama_stack.providers.inline.telemetry.meta_reference",
-            config_class="llama_stack.providers.inline.telemetry.meta_reference.config.TelemetryConfig",
-            description="Meta's reference implementation of telemetry and observability using OpenTelemetry.",
-        ),
-    ]

tests/unit/distribution/test_distribution.py

@@ -161,6 +161,8 @@ class TestProviderRegistry:
             assert internal_api not in apis, f"Internal API {internal_api} should not be in providable_apis"
 
         for api in apis:
+            if api == Api.telemetry:
+                continue
             module_name = f"llama_stack.providers.registry.{api.name.lower()}"
             try:
                 importlib.import_module(module_name)