feat(api): add prometheus monitoring (#1166)

Adds a python library for exposing basic fastapi metrics in prometheus syntax on /metrics endpoint
Add a policy to block access to the /metrics from outside the cluster
Adds a service monitor to api deployment so prometheus will scrape metrics
Redirect base url to `/docs` to prevent not found errors.

packages/api/chart/templates/istio-admin.yaml

@@ -0,0 +1,24 @@
+{{- if .Capabilities.APIVersions.Has "security.istio.io/v1beta1" }}
+apiVersion: security.istio.io/v1beta1
+kind: AuthorizationPolicy
+metadata:
+  name: api-block-metrics-access-from-public-gateway
+  namespace: {{ .Release.Namespace }}
+spec:
+  selector:
+    matchLabels:
+      {{- include "chart.selectorLabels" . | nindent 6 }}
+  action: DENY
+  rules:
+    - to:
+        - operation:
+            ports:
+              - "8080"
+            paths:
+            - /metrics*
+      from:
+        - source:
+            notNamespaces:
+            - istio-admin-gateway
+            - monitoring
+{{- end }}

packages/api/chart/templates/uds-package.yaml

@@ -7,6 +7,11 @@ metadata:
   labels:
     {{- include "chart.labels" . | nindent 4 }}
 spec:
+  monitor:
+    - portName: http
+      targetPort: {{ .Values.api.service.port }}
+      selector:
+        {{- include "chart.selectorLabels" . | nindent 8 }}
   network:
     expose:
       - service: {{ include "chart.fullname" . }}

src/leapfrogai_api/main.py

@@ -8,7 +8,7 @@
 from fastapi import FastAPI
 from fastapi.exception_handlers import request_validation_exception_handler
 from fastapi.exceptions import RequestValidationError
-
+from fastapi.responses import RedirectResponse
 from leapfrogai_api.routers.base import router as base_router
 from leapfrogai_api.routers.leapfrogai import auth
 from leapfrogai_api.routers.leapfrogai import models as lfai_models
@@ -30,6 +30,7 @@
     vector_stores,
 )
 from leapfrogai_api.utils import get_model_config
+from prometheus_fastapi_instrumentator import Instrumentator
 
 logging.basicConfig(
     level=os.getenv("LFAI_LOG_LEVEL", logging.INFO),
@@ -62,6 +63,21 @@ async def lifespan(app: FastAPI):
 app = FastAPI(lifespan=lifespan)
 
 
+@app.get("/", include_in_schema=False)
+async def root():
+    """Intercepts the root path and redirects to the API documentation."""
+    return RedirectResponse(url="/docs")
+
+
+Instrumentator(
+    excluded_handlers=["/healthz", "/metrics"],
+    should_group_status_codes=False,
+).instrument(app).expose(
+    app,
+    include_in_schema=False,
+)
+
+
 @app.exception_handler(RequestValidationError)
 async def validation_exception_handler(request, exc):
     logger.error(f"The client sent invalid data!: {exc}")

src/leapfrogai_api/pyproject.toml

@@ -26,6 +26,7 @@ dependencies = [
     "postgrest==0.16.11",                    # required by supabase, bug when using previous versions
     "openpyxl == 3.1.5",
     "psutil == 6.0.0",
+    "prometheus-fastapi-instrumentator == 7.0.0",
     "rerankers[flashrank] == 0.5.3"
 ]
 requires-python = "~=3.11"