ContextQMD
Back to Litellm

generated by https://github.com/hashicorp/terraform-plugin-docs

terraform/provider/docs/data-sources/credential.md

1.93.0-dev.14.5 KB
Original Source

litellm_credential (Data Source)

Retrieves information about an existing LiteLLM credential. This data source allows you to reference credentials that were created outside of Terraform or in other Terraform configurations.

Example Usage

terraform
# Retrieve an existing credential by name
data "litellm_credential" "existing_openai" {
  credential_name = "openai-production-key"
}

# Use the credential in a model resource
resource "litellm_model" "gpt4_with_existing_cred" {
  model_name          = "gpt-4-with-existing-cred"
  custom_llm_provider = "openai"
  base_model          = "gpt-4"
  tier                = "paid"
  mode                = "chat"
  
  # Reference the existing credential's info
  additional_litellm_params = {
    credential_name = data.litellm_credential.existing_openai.credential_name
  }
}

Example Usage with Model ID

terraform
# Retrieve a credential associated with a specific model
data "litellm_credential" "model_specific_cred" {
  credential_name = "claude-api-key"
  model_id        = "claude-3-sonnet"
}

# Use in a vector store
resource "litellm_vector_store" "knowledge_base" {
  vector_store_name       = "claude-knowledge-base"
  custom_llm_provider     = "anthropic"
  litellm_credential_name = data.litellm_credential.model_specific_cred.credential_name
  
  vector_store_description = "Knowledge base using Claude credentials"
}

Example Usage for Cross-Reference

terraform
# Get credential info to use in other resources
data "litellm_credential" "shared_cred" {
  credential_name = "shared-api-key"
}

# Create multiple resources using the same credential
resource "litellm_vector_store" "store_1" {
  vector_store_name       = "store-1"
  custom_llm_provider     = "pinecone"
  litellm_credential_name = data.litellm_credential.shared_cred.credential_name
  
  vector_store_description = "First store using shared credential"
}

resource "litellm_vector_store" "store_2" {
  vector_store_name       = "store-2"
  custom_llm_provider     = "pinecone"
  litellm_credential_name = data.litellm_credential.shared_cred.credential_name
  
  vector_store_description = "Second store using shared credential"
}

Argument Reference

The following arguments are supported:

  • credential_name - (Required) Name of the credential to retrieve.
  • model_id - (Optional) Model ID associated with this credential. Use this when the same credential name is used for different models.

Attributes Reference

In addition to all arguments above, the following attributes are exported:

  • credential_info - Map of additional non-sensitive information about the credential.

Security Note

For security reasons, the credential_values (sensitive data like API keys) are not exposed through data sources. This prevents accidental exposure of sensitive information in Terraform plans and logs. If you need to access credential values, you should manage them through the resource directly or use external secret management systems.

Common Use Cases

1. Cross-Stack References

Use data sources to reference credentials created in other Terraform configurations or stacks:

terraform
data "litellm_credential" "shared_openai" {
  credential_name = "openai-shared-key"
}

resource "litellm_model" "gpt4" {
  model_name          = "gpt-4-cross-stack"
  custom_llm_provider = "openai"
  base_model          = "gpt-4"
  
  additional_litellm_params = {
    credential_reference = data.litellm_credential.shared_openai.credential_name
  }
}

2. Conditional Logic

Use credential information for conditional resource creation:

terraform
data "litellm_credential" "optional_cred" {
  credential_name = var.credential_name
}

resource "litellm_vector_store" "conditional_store" {
  count = length(data.litellm_credential.optional_cred.credential_info) > 0 ? 1 : 0
  
  vector_store_name       = "conditional-store"
  custom_llm_provider     = "weaviate"
  litellm_credential_name = data.litellm_credential.optional_cred.credential_name
}

3. Validation and Verification

Verify that required credentials exist before creating dependent resources:

terraform
data "litellm_credential" "required_cred" {
  credential_name = "production-api-key"
}

# This will fail if the credential doesn't exist
resource "litellm_model" "production_model" {
  model_name          = "production-gpt-4"
  custom_llm_provider = "openai"
  base_model          = "gpt-4"
  
  additional_litellm_params = {
    credential_name = data.litellm_credential.required_cred.credential_name
  }
}