🎨 Adding tracing to agent and dynamic-sidecar (🏗️ DEVOPS) (ITISFo…

…undation#6691)

Co-authored-by: Andrei Neagu <[email protected]>

.env-devel

@@ -17,6 +17,7 @@ AGENT_VOLUMES_CLEANUP_S3_ENDPOINT=http://172.17.0.1:9001
 AGENT_VOLUMES_CLEANUP_S3_PROVIDER=MINIO
 AGENT_VOLUMES_CLEANUP_S3_REGION=us-east-1
 AGENT_VOLUMES_CLEANUP_S3_SECRET_KEY=12345678
+AGENT_TRACING={}
 
 API_SERVER_DEV_FEATURES_ENABLED=0
 API_SERVER_LOGLEVEL=INFO

services/agent/src/simcore_service_agent/core/application.py

@@ -5,6 +5,7 @@
     get_common_oas_options,
     override_fastapi_openapi_method,
 )
+from servicelib.fastapi.tracing import setup_tracing
 from servicelib.logging_utils import config_all_loggers
 
 from .._meta import (
@@ -59,6 +60,9 @@ def create_app() -> FastAPI:
     setup_rest_api(app)
     setup_rpc_api_routes(app)
 
+    if settings.AGENT_TRACING:
+        setup_tracing(app, settings.AGENT_TRACING, APP_NAME)
+
     async def _on_startup() -> None:
         print(APP_STARTED_BANNER_MSG, flush=True)  # noqa: T201
 

services/agent/src/simcore_service_agent/core/settings.py

@@ -6,6 +6,7 @@
 from settings_library.base import BaseCustomSettings
 from settings_library.r_clone import S3Provider
 from settings_library.rabbit import RabbitSettings
+from settings_library.tracing import TracingSettings
 from settings_library.utils_logging import MixinLoggingSettings
 
 
@@ -77,6 +78,10 @@ class ApplicationSettings(BaseCustomSettings, MixinLoggingSettings):
         auto_default_from_env=True, description="settings for service/rabbitmq"
     )
 
+    AGENT_TRACING: TracingSettings | None = Field(
+        auto_default_from_env=True, description="settings for opentelemetry tracing"
+    )
+
     @validator("LOGLEVEL")
     @classmethod
     def valid_log_level(cls, value) -> LogLevel:

services/agent/tests/conftest.py

@@ -58,6 +58,7 @@ def mock_environment(
             "RABBIT_SECURE": "false",
             "RABBIT_USER": "test",
             "AGENT_DOCKER_NODE_ID": docker_node_id,
+            "AGENT_TRACING": "null",
         },
     )
 

services/director-v2/src/simcore_service_director_v2/modules/dynamic_sidecar/docker_service_specs/sidecar.py

@@ -175,6 +175,11 @@ def _get_environment_variables(
         "S3_SECRET_KEY": r_clone_settings.R_CLONE_S3.S3_SECRET_KEY,
         "SC_BOOT_MODE": f"{app_settings.DYNAMIC_SERVICES.DYNAMIC_SIDECAR.DYNAMIC_SIDECAR_SC_BOOT_MODE}",
         "SSL_CERT_FILE": app_settings.DIRECTOR_V2_SELF_SIGNED_SSL_FILENAME,
+        "DYNAMIC_SIDECAR_TRACING": (
+            app_settings.DIRECTOR_V2_TRACING.json()
+            if app_settings.DIRECTOR_V2_TRACING
+            else "null"
+        ),
         # For background info on this special env-var above, see
         # - https://stackoverflow.com/questions/31448854/how-to-force-requests-use-the-certificates-on-my-ubuntu-system#comment78596389_37447847
         "SIMCORE_HOST_NAME": scheduler_data.service_name,

services/director-v2/tests/unit/test_modules_dynamic_sidecar_docker_service_specs_sidecar.py

@@ -37,6 +37,7 @@
     "DY_SIDECAR_USER_SERVICES_HAVE_INTERNET_ACCESS",
     "DYNAMIC_SIDECAR_COMPOSE_NAMESPACE",
     "DYNAMIC_SIDECAR_LOG_LEVEL",
+    "DYNAMIC_SIDECAR_TRACING",
     "NODE_PORTS_400_REQUEST_TIMEOUT_ATTEMPTS",
     "POSTGRES_DB",
     "POSTGRES_ENDPOINT",

services/director-v2/tests/unit/with_dbs/test_modules_dynamic_sidecar_docker_service_specs.py

@@ -243,6 +243,7 @@ def expected_dynamic_sidecar_spec(
                     "FORWARD_ENV_DISPLAY": ":0",
                     "NODE_PORTS_400_REQUEST_TIMEOUT_ATTEMPTS": "3",
                     "DYNAMIC_SIDECAR_LOG_LEVEL": "DEBUG",
+                    "DYNAMIC_SIDECAR_TRACING": "null",
                     "DY_DEPLOYMENT_REGISTRY_SETTINGS": (
                         '{"REGISTRY_AUTH": false, "REGISTRY_PATH": null, '
                         '"REGISTRY_URL": "foo.bar.com", "REGISTRY_USER": '

services/docker-compose-ops.yml

@@ -111,6 +111,7 @@ services:
       - "4318:4318"  # OTLP HTTP receiver
     networks:
       - simcore_default
+      - interactive_services_subnet
     environment:
       TRACING_OPENTELEMETRY_COLLECTOR_BATCH_SIZE: ${TRACING_OPENTELEMETRY_COLLECTOR_BATCH_SIZE}
       TRACING_OPENTELEMETRY_COLLECTOR_SAMPLING_PERCENTAGE: ${TRACING_OPENTELEMETRY_COLLECTOR_SAMPLING_PERCENTAGE}

services/docker-compose.yml

@@ -1053,6 +1053,10 @@ services:
       RABBIT_USER: ${RABBIT_USER}
       RABBIT_SECURE: ${RABBIT_SECURE}
 
+      AGENT_TRACING: ${AGENT_TRACING}
+      TRACING_OPENTELEMETRY_COLLECTOR_ENDPOINT: ${TRACING_OPENTELEMETRY_COLLECTOR_ENDPOINT}
+      TRACING_OPENTELEMETRY_COLLECTOR_PORT: ${TRACING_OPENTELEMETRY_COLLECTOR_PORT}
+
   dask-sidecar:
     image: ${DOCKER_REGISTRY:-itisfoundation}/dask-sidecar:${DOCKER_IMAGE_TAG:-latest}
     init: true

services/dynamic-sidecar/src/simcore_service_dynamic_sidecar/core/application.py

@@ -9,6 +9,7 @@
     get_common_oas_options,
     override_fastapi_openapi_method,
 )
+from servicelib.fastapi.tracing import setup_tracing
 from servicelib.logging_utils import config_all_loggers
 from simcore_sdk.node_ports_common.exceptions import NodeNotFound
 
@@ -190,6 +191,9 @@ def create_app():
     if application_settings.are_prometheus_metrics_enabled:
         setup_prometheus_metrics(app)
 
+    if application_settings.DYNAMIC_SIDECAR_TRACING:
+        setup_tracing(app, application_settings.DYNAMIC_SIDECAR_TRACING, PROJECT_NAME)
+
     # ERROR HANDLERS  ------------
     app.add_exception_handler(NodeNotFound, node_not_found_error_handler)
     app.add_exception_handler(BaseDynamicSidecarError, http_error_handler)

services/dynamic-sidecar/src/simcore_service_dynamic_sidecar/core/settings.py

@@ -23,6 +23,7 @@
 from settings_library.resource_usage_tracker import (
     DEFAULT_RESOURCE_USAGE_HEARTBEAT_INTERVAL,
 )
+from settings_library.tracing import TracingSettings
 from settings_library.utils_logging import MixinLoggingSettings
 
 
@@ -167,6 +168,10 @@ class ApplicationSettings(BaseCustomSettings, MixinLoggingSettings):
 
     SYSTEM_MONITOR_SETTINGS: SystemMonitorSettings = Field(auto_default_from_env=True)
 
+    DYNAMIC_SIDECAR_TRACING: TracingSettings | None = Field(
+        auto_default_from_env=True, description="settings for opentelemetry tracing"
+    )
+
     @property
     def are_prometheus_metrics_enabled(self) -> bool:
         return self.DY_SIDECAR_CALLBACKS_MAPPING.metrics is not None

services/dynamic-sidecar/tests/conftest.py

@@ -199,6 +199,7 @@ def base_mock_envs(
                 "REGISTRY_SSL": "false",
             }
         ),
+        "DYNAMIC_SIDECAR_TRACING": "null",
     }