👥📊 Team/Key Based Logging

Allow each key/team to use their own Langfuse Project / custom callbacks

This allows you to do the following

Team 1 -> Logs to Langfuse Project 1 
Team 2 -> Logs to Langfuse Project 2
Team 3 -> Disabled Logging (for GDPR compliance)

Team Based Logging

👉 Tutorial - Allow each team to use their own Langfuse Project / custom callbacks

Logging / Caching

Turn on/off logging and caching for a specific team id.

Example:

This config would send langfuse logs to 2 different langfuse projects, based on the team id

litellm_settings:
  default_team_settings: 
    - team_id: "dbe2f686-a686-4896-864a-4c3924458709"
      success_callback: ["langfuse"]
      langfuse_public_key: os.environ/LANGFUSE_PUB_KEY_1 # Project 1
      langfuse_secret: os.environ/LANGFUSE_PRIVATE_KEY_1 # Project 1
    - team_id: "06ed1e01-3fa7-4b9e-95bc-f2e59b74f3a8"
      success_callback: ["langfuse"]
      langfuse_public_key: os.environ/LANGFUSE_PUB_KEY_2 # Project 2
      langfuse_secret: os.environ/LANGFUSE_SECRET_2 # Project 2

Now, when you generate keys for this team-id

curl -X POST 'http://0.0.0.0:4000/key/generate' \
-H 'Authorization: Bearer sk-1234' \
-H 'Content-Type: application/json' \
-d '{"team_id": "06ed1e01-3fa7-4b9e-95bc-f2e59b74f3a8"}'

All requests made with these keys will log data to their team-specific logging. -->

[BETA] Team Logging via API

info

✨ This is an Enterprise only feature Get Started with Enterprise here

Set Callbacks Per Team

1. Set callback for team

We make a request to POST /team/{team_id}/callback to add a callback for

curl -X POST 'http:/localhost:4000/team/dbe2f686-a686-4896-864a-4c3924458709/callback' \
-H 'Content-Type: application/json' \
-H 'Authorization: Bearer sk-1234' \
-d '{
  "callback_name": "langfuse",
  "callback_type": "success",
  "callback_vars": {
    "langfuse_public_key": "pk", 
    "langfuse_secret_key": "sk_", 
    "langfuse_host": "https://cloud.langfuse.com"
    }
  
}'

Supported Values

Field	Supported Values	Notes
callback_name	"langfuse", "gcs_bucket"	Currently only supports "langfuse", "gcs_bucket"
callback_type	"success", "failure", "success_and_failure"
callback_vars		dict of callback settings
langfuse_public_key	string	Required for Langfuse
langfuse_secret_key	string	Required for Langfuse
langfuse_host	string	Optional for Langfuse (defaults to https://cloud.langfuse.com)
gcs_bucket_name	string	Required for GCS Bucket. Name of your GCS bucket
gcs_path_service_account	string	Required for GCS Bucket. Path to your service account json

2. Create key for team

All keys created for team dbe2f686-a686-4896-864a-4c3924458709 will log to langfuse project specified on Step 1. Set callback for team

curl --location 'http://0.0.0.0:4000/key/generate' \
    --header 'Authorization: Bearer sk-1234' \
    --header 'Content-Type: application/json' \
    --data '{
        "team_id": "dbe2f686-a686-4896-864a-4c3924458709"
}'

3. Make /chat/completion request for team

curl -i http://localhost:4000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-KbUuE0WNptC0jXapyMmLBA" \
  -d '{
    "model": "gpt-4",
    "messages": [
      {"role": "user", "content": "Hello, Claude gm!"}
    ]
}'

Expect this to be logged on the langfuse project specified on Step 1. Set callback for team

Disable Logging for a Team

To disable logging for a specific team, you can use the following endpoint:

POST /team/{team_id}/disable_logging

This endpoint removes all success and failure callbacks for the specified team, effectively disabling logging.

Step 1. Disable logging for team

curl -X POST 'http://localhost:4000/team/YOUR_TEAM_ID/disable_logging' \
    -H 'Authorization: Bearer YOUR_API_KEY'

Replace YOUR_TEAM_ID with the actual team ID

Response A successful request will return a response similar to this:

{
    "status": "success",
    "message": "Logging disabled for team YOUR_TEAM_ID",
    "data": {
        "team_id": "YOUR_TEAM_ID",
        "success_callbacks": [],
        "failure_callbacks": []
    }
}

Step 2. Test it - /chat/completions

Use a key generated for team = team_id - you should see no logs on your configured success callback (eg. Langfuse)

curl -i http://localhost:4000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-KbUuE0WNptC0jXapyMmLBA" \
  -d '{
    "model": "gpt-4",
    "messages": [
      {"role": "user", "content": "Hello, Claude gm!"}
    ]
}'

Debugging / Troubleshooting

• Check active callbacks for team using GET /team/{team_id}/callback

Use this to check what success/failure callbacks are active for team=team_id

curl -X GET 'http://localhost:4000/team/dbe2f686-a686-4896-864a-4c3924458709/callback' \
        -H 'Authorization: Bearer sk-1234'

Team Logging Endpoints

• POST /team/{team_id}/callback Add a success/failure callback to a team
• GET /team/{team_id}/callback - Get the success/failure callbacks and variables for a team

[BETA] Key Based Logging

Use the /key/generate or /key/update endpoints to add logging callbacks to a specific key.

info

✨ This is an Enterprise only feature Get Started with Enterprise here

How key based logging works:

• If Key has no callbacks configured, it will use the default callbacks specified in the config.yaml file
• If Key has callbacks configured, it will use the callbacks specified in the key

Langfuse

curl -X POST 'http://0.0.0.0:4000/key/generate' \
-H 'Authorization: Bearer sk-1234' \
-H 'Content-Type: application/json' \
-d '{
    "metadata": {
        "logging": [{
            "callback_name": "langfuse", # "otel", "gcs_bucket"
            "callback_type": "success", # "success", "failure", "success_and_failure"
            "callback_vars": {
                "langfuse_public_key": "os.environ/LANGFUSE_PUBLIC_KEY", # [RECOMMENDED] reference key in proxy environment
                "langfuse_secret_key": "os.environ/LANGFUSE_SECRET_KEY", # [RECOMMENDED] reference key in proxy environment
                "langfuse_host": "https://cloud.langfuse.com"
            }
        }]
    }
}'

GCS Bucket

1. Create Virtual Key to log to a specific GCS Bucket

   Set GCS_SERVICE_ACCOUNT in your environment to the path of the service account json

   export GCS_SERVICE_ACCOUNT=/path/to/service-account.json # GCS_SERVICE_ACCOUNT=/Users/ishaanjaffer/Downloads/adroit-crow-413218-a956eef1a2a8.json

   curl -X POST 'http://0.0.0.0:4000/key/generate' \
   -H 'Authorization: Bearer sk-1234' \
   -H 'Content-Type: application/json' \
   -d '{
       "metadata": {
           "logging": [{
               "callback_name": "gcs_bucket", # "otel", "gcs_bucket"
               "callback_type": "success", # "success", "failure", "success_and_failure"
               "callback_vars": {
                   "gcs_bucket_name": "my-gcs-bucket", # Name of your GCS Bucket to log to
                   "gcs_path_service_account": "os.environ/GCS_SERVICE_ACCOUNT" # environ variable for this service account
               }
           }]
       }
   }'
   

2. Test it - /chat/completions request

   Use the virtual key from step 3 to make a /chat/completions request

   You should see your logs on GCS Bucket on a successful request

   curl -i http://localhost:4000/v1/chat/completions \
     -H "Content-Type: application/json" \
     -H "Authorization: Bearer sk-Fxq5XSyWKeXDKfPdqXZhPg" \
     -d '{
       "model": "fake-openai-endpoint",
       "messages": [
         {"role": "user", "content": "Hello, Claude"}
       ],
       "user": "hello",
     }'

---

Help us improve this feature, by filing a ticket here

Check if key callbacks are configured correctly /key/health

Call /key/health with the key to check if the callback settings are configured correctly

Pass the key in the request header

curl -X POST "http://localhost:4000/key/health" \
  -H "Authorization: Bearer <your-key>" \
  -H "Content-Type: application/json"

Response when key is configured correctly

Response when logging callbacks are setup correctly:

A key is healthy when the logging callbacks are setup correctly.

{
  "key": "healthy",
  "logging_callbacks": {
    "callbacks": [
      "gcs_bucket"
    ],
    "status": "healthy",
    "details": "No logger exceptions triggered, system is healthy. Manually check if logs were sent to ['gcs_bucket']"
  }
}

Response when key is configured incorrectly

Response when logging callbacks are not setup correctly

A key is unhealthy when the logging callbacks are not setup correctly.

{
  "key": "unhealthy",
  "logging_callbacks": {
    "callbacks": [
      "gcs_bucket"
    ],
    "status": "unhealthy",
    "details": "Logger exceptions triggered, system is unhealthy: Failed to load vertex credentials. Check to see if credentials containing partial/invalid information."
  }
}