Add OpenTelemetry Config (#2194)

* Add OpenTelemetry Config

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Added psycopg

* Fix pre-commit test issues

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

Author: shaidar

main/apps.py

@@ -14,3 +14,8 @@ def ready(self):
         from main import features
 
         features.configure()
+
+        # Initialize OpenTelemetry
+        from main.telemetry import configure_opentelemetry
+
+        configure_opentelemetry()

main/settings.py

@@ -833,3 +833,11 @@ def get_all_config_keys():
 }
 
 CONTENT_FILE_SUMMARIZER_BATCH_SIZE = get_int("CONTENT_FILE_SUMMARIZER_BATCH_SIZE", 20)
+
+# OpenTelemetry configuration
+OPENTELEMETRY_ENABLED = get_bool("OPENTELEMETRY_ENABLED", False)  # noqa: FBT003
+OPENTELEMETRY_SERVICE_NAME = get_string("OPENTELEMETRY_SERVICE_NAME", "learn")
+OPENTELEMETRY_INSECURE = get_bool("OPENTELEMETRY_INSECURE", default=True)
+OPENTELEMETRY_ENDPOINT = get_string("OPENTELEMETRY_ENDPOINT", None)
+OPENTELEMETRY_TRACES_BATCH_SIZE = get_int("OPENTELEMETRY_TRACES_BATCH_SIZE", 512)
+OPENTELEMETRY_EXPORT_TIMEOUT_MS = get_int("OPENTELEMETRY_EXPORT_TIMEOUT_MS", 5000)

main/telemetry.py

@@ -0,0 +1,87 @@
+"""OpenTelemetry initialization and configuration for Learn AI."""
+
+import logging
+from typing import Optional
+
+from django.conf import settings
+from opentelemetry import trace
+from opentelemetry.exporter.otlp.proto.grpc.trace_exporter import OTLPSpanExporter
+from opentelemetry.instrumentation.celery import CeleryInstrumentor
+from opentelemetry.instrumentation.django import DjangoInstrumentor
+from opentelemetry.instrumentation.psycopg import PsycopgInstrumentor
+from opentelemetry.instrumentation.redis import RedisInstrumentor
+from opentelemetry.instrumentation.requests import RequestsInstrumentor
+from opentelemetry.sdk.resources import Resource
+from opentelemetry.sdk.trace import TracerProvider
+from opentelemetry.sdk.trace.export import BatchSpanProcessor, ConsoleSpanExporter
+
+log = logging.getLogger(__name__)
+
+
+def configure_opentelemetry() -> Optional[TracerProvider]:
+    """
+    Configure OpenTelemetry with appropriate instrumentations and exporters.
+    Returns the tracer provider if configured, None otherwise.
+    """
+    if not getattr(settings, "OPENTELEMETRY_ENABLED", False):
+        log.info("OpenTelemetry is disabled")
+        return None
+
+    log.info("Initializing OpenTelemetry")
+
+    # Create a resource with service info
+    resource = Resource.create(
+        {
+            "service.name": getattr(settings, "OPENTELEMETRY_SERVICE_NAME", "learn"),
+            "service.version": getattr(settings, "VERSION", "unknown"),
+            "deployment.environment": settings.ENVIRONMENT,
+        }
+    )
+
+    # Configure the tracer provider
+    tracer_provider = TracerProvider(resource=resource)
+    trace.set_tracer_provider(tracer_provider)
+
+    # Add console exporter for development/testing
+    if settings.DEBUG:
+        log.info("Adding console exporter for OpenTelemetry")
+        console_exporter = ConsoleSpanExporter()
+        tracer_provider.add_span_processor(BatchSpanProcessor(console_exporter))
+
+    # Add OTLP exporter if configured
+    otlp_endpoint = getattr(settings, "OPENTELEMETRY_ENDPOINT", None)
+    if otlp_endpoint:
+        log.info("Configuring OTLP exporter to endpoint: %s", otlp_endpoint)
+
+        headers = {}
+
+        try:
+            otlp_exporter = OTLPSpanExporter(
+                endpoint=otlp_endpoint,
+                headers=headers,
+                insecure=getattr(settings, "OPENTELEMETRY_INSECURE", True),
+            )
+
+            tracer_provider.add_span_processor(
+                BatchSpanProcessor(
+                    otlp_exporter,
+                    max_export_batch_size=getattr(
+                        settings, "OPENTELEMETRY_BATCH_SIZE", 512
+                    ),
+                    schedule_delay_millis=getattr(
+                        settings, "OPENTELEMETRY_EXPORT_TIMEOUT_MS", 5000
+                    ),
+                )
+            )
+        except Exception:
+            log.exception("Failed to configure OTLP exporter")
+
+    # Initialize instrumentations
+    DjangoInstrumentor().instrument()
+    PsycopgInstrumentor().instrument()
+    RedisInstrumentor().instrument()
+    CeleryInstrumentor().instrument()
+    RequestsInstrumentor().instrument()
+
+    log.info("OpenTelemetry initialized successfully")
+    return tracer_provider