fix: Add opentelemtry-based tracing

This adds tracing to fastapi, redis, and sqlalchemy. It uses the
recommended OLTP exporter to send the tracing data.

aurweb/asgi.py

@@ -14,6 +14,12 @@ from fastapi import FastAPI, HTTPException, Request, Response
 from fastapi.responses import RedirectResponse
 from fastapi.staticfiles import StaticFiles
 from jinja2 import TemplateNotFound
+from opentelemetry import trace
+from opentelemetry.exporter.otlp.proto.http.trace_exporter import OTLPSpanExporter
+from opentelemetry.instrumentation.fastapi import FastAPIInstrumentor
+from opentelemetry.sdk.resources import Resource
+from opentelemetry.sdk.trace import TracerProvider
+from opentelemetry.sdk.trace.export import BatchSpanProcessor
 from sqlalchemy import and_
 from starlette.exceptions import HTTPException as StarletteHTTPException
 from starlette.middleware.authentication import AuthenticationMiddleware
@@ -54,6 +60,17 @@ instrumentator().add(prometheus.http_requests_total())
 instrumentator().instrument(app)
 
 
+# Instrument FastAPI for tracing
+FastAPIInstrumentor.instrument_app(app)
+
+resource = Resource(attributes={"service.name": "aurweb"})
+otlp_endpoint = aurweb.config.get("tracing", "otlp_endpont")
+otlp_exporter = OTLPSpanExporter(endpoint=otlp_endpoint, insecure=True)
+span_processor = BatchSpanProcessor(otlp_exporter)
+trace.set_tracer_provider(TracerProvider(resource=resource))
+trace.get_tracer_provider().add_span_processor(span_processor)
+
+
 async def app_startup():
     # https://stackoverflow.com/questions/67054759/about-the-maximum-recursion-error-in-fastapi
     # Test failures have been observed by internal starlette code when

aurweb/aur_redis.py

@@ -1,4 +1,5 @@
 import fakeredis
+from opentelemetry.instrumentation.redis import RedisInstrumentor
 from redis import ConnectionPool, Redis
 
 import aurweb.config
@@ -7,6 +8,8 @@ from aurweb import aur_logging
 logger = aur_logging.get_logger(__name__)
 pool = None
 
+RedisInstrumentor().instrument()
+
 
 class FakeConnectionPool:
     """A fake ConnectionPool class which holds an internal reference

aurweb/db.py

@@ -298,9 +298,12 @@ def get_engine(dbname: str = None, echo: bool = False):
             connect_args["check_same_thread"] = False
 
         kwargs = {"echo": echo, "connect_args": connect_args}
+        from opentelemetry.instrumentation.sqlalchemy import SQLAlchemyInstrumentor
         from sqlalchemy import create_engine
 
-        _engines[dbname] = create_engine(get_sqlalchemy_url(), **kwargs)
+        engine = create_engine(get_sqlalchemy_url(), **kwargs)
+        SQLAlchemyInstrumentor().instrument(engine=engine)
+        _engines[dbname] = engine
 
         if is_sqlite:  # pragma: no cover
             setup_sqlite(_engines.get(dbname))

conf/config.defaults

@@ -177,3 +177,6 @@ expiry_time_search = 600
 expiry_time_statistics = 300
 ; number of seconds after a cache entry for rss queries expires, default is 5 minutes
 expiry_time_rss = 300
+
+[tracing]
+otlp_nedpoint = "http://localhost:4318"

pyproject.toml

@@ -69,7 +69,6 @@ httpx = "^0.27.0"
 itsdangerous = "^2.1.2"
 lxml = "^5.2.1"
 orjson = "^3.10.0"
-protobuf = "^5.26.1"
 pygit2 = "^1.14.1"
 python-multipart = "^0.0.9"
 redis = "^5.0.3"
@@ -89,7 +88,6 @@ SQLAlchemy = "^1.4.52"
 uvicorn = "^0.29.0"
 gunicorn = "^21.2.0"
 Hypercorn = "^0.16.0"
-prometheus-fastapi-instrumentator = "^7.0.0"
 pytest-xdist = "^3.5.0"
 filelock = "^3.13.3"
 posix-ipc = "^1.1.1"
@@ -98,6 +96,15 @@ fastapi = "^0.110.0"
 srcinfo = "^0.1.2"
 tomlkit = "^0.12.0"
 
+# Tracing
+prometheus-fastapi-instrumentator = "^7.0.0"
+opentelemetry-api = "^1.26.0"
+opentelemetry-sdk = "^1.26.0"
+opentelemetry-exporter-otlp-proto-http = "^1.26.0"
+opentelemetry-instrumentation-fastapi = "^0.47b0"
+opentelemetry-instrumentation-redis = "^0.47b0"
+opentelemetry-instrumentation-sqlalchemy = "^0.47b0"
+
 [tool.poetry.dev-dependencies]
 coverage = "^7.4.4"
 pytest = "^8.1.1"