Fix Langfuse callback and Traefik health check issues

- Switch from langfuse_otel to native langfuse callback (OTEL endpoint not available)
- Change LANGFUSE_OTEL_HOST to LANGFUSE_HOST env var for native SDK
- Update callback_settings from otel to langfuse
- Add Traefik health check labels pointing to /health/liveliness endpoint
  to prevent recurring auth errors from unauthenticated health probes

config/general_settings.yaml

@@ -35,5 +35,5 @@ general_settings:
   use_client_credentials_pass_through_routes: true  # use client credentials for all pass through routes like "/vertex-ai", /bedrock/. When this is True Virtual Key auth will not be applied on these endpoints
 
 callback_settings:
-  otel:
-    message_logging: true  # OTEL logging callback specific settings
+  langfuse:
+    message_logging: true  # Langfuse logging callback specific settings

config/litellm_settings.yaml

@@ -1,9 +1,7 @@
 litellm_settings:
   # Logging/Callback settings
-  success_callback: ["langfuse_otel"]  # list of success callbacks
-  failure_callback: ["otel"]  # list of failure callbacks
-  # callbacks: ["otel"]  # list of callbacks - runs on success and failure
-  # service_callbacks: ["datadog", "prometheus"]  # logs redis, postgres failures on datadog, prometheus
+  callbacks: ["langfuse"]  # runs on success and failure (native SDK, not OTEL)
+  # service_callbacks: ["datadog", "prometheus"]  # logs redis, postgres failures on datadog, prometheus (enterprise only)
   turn_off_message_logging: false  # prevent the messages and responses from being logged to on your callbacks, but request metadata will still be logged. Useful for privacy/compliance when handling sensitive data.
   redact_user_api_key_info: false  # Redact information about the user api key (hashed token, user_id, team id, etc.), from logs. Currently supported for Langfuse, OpenTelemetry, Logfire, ArizeAI logging.
   langfuse_default_tags: ["cache_hit", "cache_key", "proxy_base_url", "user_api_key_alias", "user_api_key_user_id", "user_api_key_user_email", "user_api_key_team_alias", "semantic-similarity", "proxy_base_url"] # default tags for Langfuse Logging

docker-compose.yml

@@ -63,6 +63,10 @@ services:
                 - "traefik.http.routers.litellm.rule=Host(`llm.lab`) || Host(`llm.toy`)"
                 - "traefik.http.routers.litellm.service=litellm"
                 - "traefik.http.services.litellm.loadbalancer.server.port=4000"
+                # Health check using unauthenticated endpoint
+                - "traefik.http.services.litellm.loadbalancer.healthcheck.path=/health/liveliness"
+                - "traefik.http.services.litellm.loadbalancer.healthcheck.interval=30s"
+                - "traefik.http.services.litellm.loadbalancer.healthcheck.timeout=5s"
               
 networks:
     net:

stack.env

@@ -8,7 +8,7 @@
 DATABASE_URL=postgresql://vasceannie:squirtle123456@192.168.50.210:5432/sandbox?schema=pvgo7i70amd9xkx
 LANGFUSE_PUBLIC_KEY=pk-lf-d19aa5c3-9bd2-4cda-953a-74d67d174602
 LANGFUSE_SECRET_KEY=sk-lf-dd3a0363-2d85-4291-9a42-91d5f8181a1a
-LANGFUSE_OTEL_HOST=http://fuse.lab
+LANGFUSE_HOST=http://fuse.lab
 UI_USERNAME=vasceannie
 UI_PASSWORD=$QUirtle123
 LITELLM_MASTER_KEY=sk-1234