⚙️ Airflow Copilot – Environment Configuration

This document describes all the environment variables required for running Airflow Copilot, including their purpose, default values, and behavior across different authentication strategies.

---

🧠 Core Configuration

Variable	Description	Required	Default
LLM_MODEL_PROVIDER_NAME	The name of the LLM provider (e.g., openai, google_genai)	✅	–
LLM_MODEL_NAME	The specific model to use (e.g., gpt-4, mixtral-8x7b)	✅	–
{PROVIDER}_API_KEY	API key used to authenticate with the selected LLM provider	✅	–
DB_URI	PostgreSQL URI for persisting agent state and user data	✅	–
MICROSOFT_APP_ID	Azure Bot App registration ID	✅	–
MICROSOFT_APP_PASSWORD	Secret/password for the registered bot	✅	–

---

• {PROVIDER}_API_KEY- Replace the {PROVIDER} with the LLM_MODEL_PROVIDER_NAME (in Uppercase). For example , OPENAI_API_KEY or GOOGLE_GENAI_API_KEY etc.

🧾 Summarization Configuration

Variable	Description	Required	Default
SUMMARIZATION_LLM_MODEL_PROVIDER_NAME	LLM provider to use specifically for summarization (fallbacks to core provider if unset)	❌	Provider above
SUMMARIZATION_LLM_MODEL_NAME	LLM model name for summarization (fallbacks to core model if unset)	❌	Model above
SUMMARIZATION_LLM_MODEL_KEY	LLM model key for summarization (fallbacks to core model if unset)	❌	above key
MIN_MSG_TO_RETAIN	Minimum messages to keep before summarizing earlier messages	✅	5
MIN_MSG_TO_SUMMARIZE	Minimum messages required (beyond retained) to trigger summarization	✅	5

---

🔐 Airflow Authentication

Variable	Description	Required	Default
AIRFLOW_AUTH_STRATEGY	Defines credential strategy: centralized or per_user	✅	per_user

Centralized Strategy

Variable	Description	Required
AIRFLOW_BASE_URL	Base URL of your Apache Airflow instance	✅
AIRFLOW_USER_NAME	Global user for calling Airflow APIs	✅
AIRFLOW_USER_PASSWORD	Password for the global Airflow user	✅

Per User Strategy

Variable	Description	Required
FERNET_SECRET_KEY	Secret key used for encrypting user credentials in the DB	✅

Variable for Local Deployment

Variable	Description	Required
AZURE_BOT_NAME	Name of the Bot that you have created. e.g Airflow-Copilot	❌
NGROK_AUTHTOKEN	NGROK Auth Token	✅

---

🔍 Environment Variable Details

1. LLM_MODEL_PROVIDER_NAME

   • Description: Name of the LLM provider responsible for AI/agent interactions.
   • Default: None
   • Possible values: openai, google_genai, anthropic.
   • Required: ✅ Yes

2. LLM_MODEL_NAME

   • Description: The LLM model used for primary reasoning and responses.
   • Default: None
   • Possible values: gpt-4, mixtral-8x7b, gemini-pro, etc.
   • Required: ✅ Yes

3. {PROVIDER}_API_KEY

   • Description: API key used to authenticate with the selected LLM provider. Replace {PROVIDER} with the uppercase value of LLM_MODEL_PROVIDER_NAME.
   • Default: None
   • Possible values: Varies by provider
   • Required:

4. DB_URI
   - Description: PostgreSQL URI used for persisting state, conversation checkpoints, and user credentials.
   - Default: None
   - Possible values: e.g., 'postgresql://username:password@host:port/db'
   - Required: ✅ Yes

5. MICROSOFT_APP_ID

   • Description: Azure Bot registration ID. Required to connect your bot to Microsoft Teams.
   • Default: None
   • Possible values: Azure-generated GUID
   • Required: ✅ Yes

6. MICROSOFT_APP_PASSWORD

   • Description: Secret associated with the Azure Bot App ID.
   • Default: None
   • Possible values: Azure-generated client secret
   • Required: ✅ Yes

7. SUMMARIZATION_LLM_MODEL_PROVIDER_NAME

   • Description: Optional override for the LLM provider used in summarization. Falls back to the main provider if not set.
   • Default: value of LLM_MODEL_PROVIDER_NAME
   • Possible values: Same as LLM providers
   • Required: ❌ Optional

8. SUMMARIZATION_LLM_MODEL_NAME

   • Description: Optional override for the model used in summarization. Falls back to main model if not set.
   • Default: value of LLM_MODEL_NAME
   • Possible values: Same as other models
   • Required: ❌ Optional

9. MIN_MSG_TO_RETAIN

   • Description: Number of most recent user/tool messages to retain in memory before summarization.
   • Default: 10
   • Possible values: Integer > 0
   • Required: ✅ Yes

10. MIN_MSG_TO_SUMMARIZE

    • Description: Minimum number of messages beyond retained ones to trigger summarization.
    • Default: 10
    • Possible values: Integer > 0
    • Required: ✅ Yes

11. AIRFLOW_AUTH_STRATEGY

    • Description: Defines how Airflow credentials are managed – centrally or per user.
    • Default: per_user
    • Possible values: centralized, per_user
    • Required: ✅ Yes

12. AIRFLOW_BASE_URL

    • Description: Base URL of the Airflow instance used for API calls.
    • Default: None
    • Possible values: e.g., https://<host>:<port>
    • Required: ✅ Yes

13. AIRFLOW_USER_NAME

    • Description: Username for Airflow (only for centralized strategy).
    • Default: None
    • Possible values: Any valid Airflow user
    • Required: ✅ Yes (if centralized)

14. AIRFLOW_USER_PASSWORD

    • Description: Password for Airflow user (only for centralized strategy).
    • Default: None
    • Possible values: Any valid password
    • Required: ✅ Yes (if centralized)

15. FERNET_SECRET_KEY

    • Description: Encryption key used to store user-specific Airflow credentials in the database.
    • Default: None
    • Possible values: Base64-encoded 32-byte key
    • Required: ✅ Yes (if per_user)
    • Command: Run below command to generate the key.

    python -c "from cryptography.fernet import Fernet; print(Fernet.generate_key().decode())"
    

    Do not change the Farnet key in every deployment otherwise the backend credentials will not be decrypt and every user have to update their cred again.

16. AZURE_BOT_NAME

    • Description: The Azure Bot Name to automatically update the endpoint while Local Deployment.
    • Default: Airflow-Copilot
    • Required: ❌ Optional

17. NGROK_AUTHTOKEN

    • Description: The Ngrok Token to deploy the Local Fast API to public Https endpoint.
    • Required: ✅ Yes (for Local deployment)