Deployment

The deployment process follows a sequence that builds upon previous steps, ensuring each component is appropriately configured before proceeding to the next phase.

1: Environment Preparation

Procedure

  1. Create a dedicated namespace for MaxAI to provide resource isolation and easier management. Run the following command to establish a dedicated environment: 
    kubectl create namespace maxai
  2. Run the following command to configure the kubectl context to use the MaxAI namespace by default. It prevents accidental deployment of resources to the wrong namespace. 
    kubectl config set-context --current --namespace=maxai

2: Customizing Configurations

Procedure

  1. Create a custom values file by copying the default values.yaml file and modifying it for your specific environment.
  2. Name this file something descriptive like maxai-production-values.yaml to clearly identify its purpose.

3: Configuring the values.yaml File

About this task

Before installing the Helm charts, configure the downloaded values.yaml file as per your environment settings:

Procedure

  1. From the deployment bundle, open the values.yaml file.
  2. Configure the following parameters as per your environment setup:
    Note: Some parameters should not be modified, some parameters do not need to be configured, and some parameters need to be configured. The details for the parameters are mentioned in the following table:
    Parameter Name Type Description and Example Values
    hclImagePullSecret To be configured Example: regcred
    mariadb_ip To be configured The MariaDB hostname or IP address.

    Example: maxai-suite-database
    mariadb_port To be configured The port number of the MariaDB hostname or IP address.

    Example: 3306
    securityContext To be configured The values for both sub-parameters must be the same:

    Example:

    • runAsUser - 5002
    • runAsGroup - 5002
    BASIC_AUTH_USERNAME To be configured The authorization credentials.

    Example: admin
    BASIC_AUTH_PASSWORD To be configured The authorization credentials.

    Example: secret
    CIPHER_KEY Not Applicable
    MODEL_ACCESS_KEY Not Applicable
    TOKEN_EXEMPTION_USERS To be configured Configure the users that you want to be exempt from GenAI token limitation. You can configure multiple users.

    Example: username1,username2
    PLATFORM_BASE_URL To be configured HCL Unica Platform URL or service name (if MaxAI is deployed on same network)

    Example: http://unica-platform-domain:1111
    MAX_AI_CONFIG_PATH To be configured The path where MaxAI configurations are implemented on HCL Unica.

    Example: Affinium|AION
    CLIENT_ID To be configured Credentials for accessing HCL Unica Platform.

    Example: username
    ACCESS_KEY To be configured Credentials for accessing HCL Unica Platform.

    Example: password
    chatBotServerURL Do not change Example: CHATBOT_HOST_URL
    HOST Do not change Example: localhost
    PORT Do not change Example: 9000
    CONTEXT_ROOT Do not change Example: /platform
    REDIS_HOST To be configured The hostname of the Redis service.

    Example: redis-leader
    REDIS_PORT To be configured The port number for the Redis service.

    Example: 6379
    REDIS_PASSWORD To be configured The password for Redis service.

    Example: password
    REDIS_DB Do not change Example: 0
    REDIS_SESSION_TTL Do not change Example: 3600
    VERIFY_SSL To be configured Determines whether SSL certificates should be verified when making HTTPS requests. On Production is should be true.

    Example: true
    SECURE_COOKIE To be configured Specifies whether cookies should be marked as secured (transmitted only over HTTPS). On Production is should be true.

    Example: true
    TOKEN_EXEMPT_SERVICES To be configured Configure the services that you want to be exempt from GenAI token limitation. You can configure either the email-generation service, the sla service or both.

    Example: email-generation,sla
    API_REQUEST_TIMEOUT To be configured The API request timeout, in seconds, for analytics service.

    Example: 300
    ANALYTICS Do not change The service name for Analytics.

    Example: maxai-analytics
    INSIGHTS Do not change The service name for Insights.

    Example: maxai-insights
    OFFER Do not change The service name for Offer.

    Example: maxai-offer
    CLASSIFICATION Do not change The service name for Classification Engine.

    Example: maxai-classification-engine
    EMAIL Do not change The service name for Email Generation.

    Example: maxai-email-generation
    CONFIG Do not change The service name for Config Service.

    Example: maxai-config-service
    SLA Do not change The service name for SLA.

    Example: maxai-sla-service
    DOCUMENT_SEARCH Do not change The service name for Document search.

    Example: maxai-document-search
    API_GATEWAY Do not change The service name for API Gateway.

    Example: maxai-analytics
    LOG_LEVEL Do not change Used for configuring logger.

    Example: DEBUG
    LOG_TO_FILE Do not change Used for configuring logger.

    Example: False
    LOG_FILE Do not change Used for configuring logger.

    Example: logs/app.log
    LOG_FILE_ROTATION We recommend that you do not change. Can be modified, if required. Example: 10 MB
    OLD_LOG_FILES_RETENTION Not Applicable
    OLD_LOG_FILES_COMPRESSION Not Applicable
    MAX_HISTORY Do not change Example: 2
    MODERATION_API Do not change Example: https://api.openai.com/v1/moderations
    CONST_SQL_API_TIMEOUT Do not change Example: 200
    FEEDBACK_LOG_PATH We recommend that you do not change. Can be modified, if required. The path where the session logs are stored.

    Example: /mnt/hclgenai_share/ce/chat_session_logs
    OPENAI_ASSISTANT_ID To be configured OpenAI RAG assistant for docsearch.

    Example: assistant_XxXXXXXXxxXX0xX2XxX0X00
    USE_ASSISTANT To be configured Enable or disable OpenAI RAG assistant. Disabled by default.

4: Installing Helm Charts

Procedure

  1. Navigate to the directory containing the MaxAI chart.
  2. Execute the following Helm installation command: 
    helm install genai ./hclgenai-hclnow -f maxai-production-values.yaml --namespace maxai
    • The installation name genai can be customized based on your naming conventions.
    • The directory name hclgenai-hclnow has to be the directory name where your deployment bundle is present.
    • The maxai-production-values.yaml is the values.yaml file, customized for your environment.
    • The namespace maxai is the one you have created for deploying the MaxAI suite of applications.
    Run the following command to monitor the installation progress and to watch pods start up in real-time:
    kubectl get pods -w

    The deployment process typically takes several minutes as containers are pulled and started.

    Change the permission for the shared mount point to genaiuser with uid and gid as 5002.
    chown 5002:5002 /mnt/hclgenai_share

5: Run the Provided SQL scripts

About this task

After installation and before verification, download and run the SQL script on MaxAI database ce_metering_db. Remember to run the script before moving on to verification. Download Script.

6: Initial Verification

Procedure

  1. If the installation is successful, all the pods will be up and running successfully. A Sample deployment output follows: 
    
NAME                                                     READY   STATUS      RESTARTS   AGE
genai-hcl-genai-suite-database-0                         1/1     Running     0          7h25m
genai-hclgenai-classification-engine-544fdcdf6b-n7hhn    1/1     Running     0          7h25m
genai-hclgenai-documentsearch-6cd78484c9-d9jgv           1/1     Running     0          7h25m
genai-hclgenai-email-gen-64b5f8cf54-zjrfr                1/1     Running     0          7h25m
genai-hclgenai-maxai-analytics-5c95cfc5f8-zcjnq          1/1     Running     0          7h25m
genai-hclgenai-maxai-insights-6666c45568-7hwzn           1/1     Running     0          7h25m
genai-hclgenai-sla-56786f8d68-phpbn                      1/1     Running     0          7h25m
genai-maxai-api-gateway-7f4cc44788-snqwk                 1/1     Running     0          7h25m
genai-maxai-config-service-5867d7d6d6-4l6zw              1/1     Running     0          7h25m
genai-maxai-offer-6b9788c7cf-jzrjf                       1/1     Running     0          7h25m
redis-leader-858cb8c796-kdp8p                            1/1     Running     0          7h25m
  2. You can also check service status to ensure networking is configured correctly and that load balancers or ingress controllers have been properly provisioned. Run the following command: 
    kubectl get services

What to do next

After deploying MaxAI, you must configure it. For information related to MaxAI configurations, see HCL MaxAI Administrator's Guide.