Configure BYO AI In Mini DBA


Bring Your Own AI, or BYO AI, lets you use your own provider subscription for Mini DBA AI Assistant requests. This is useful when your organization requires an approved model, private endpoint, specific region, local model, or centralized billing.

Mini DBA BYO AI settings screenshot placeholder

Why Use BYO AI

The default Mini DBA AI provider can be useful for quick analysis, but it has a daily free token allowance. BYO AI sends AI Assistant requests to the provider you configure in AI Provider Settings. That means usage is governed by your provider subscription, model limits, network access, and internal policy rather than the Mini DBA default free token quota.

Use BYO AI when you need:

  • A company-approved AI model or provider.
  • Azure OpenAI with Entra authentication or a private Azure endpoint.
  • An OpenAI-compatible gateway, model router, or hosted inference endpoint.
  • Local or self-hosted inference through Ollama.
  • Anthropic Claude for console AI requests.
  • Separate cost tracking and provider-side logging.

Supported Provider Choices

Provider Typical Use Authentication Options
OpenAI Standard OpenAI API usage. API key.
Azure OpenAI Enterprise Azure deployment and regional control. API key or Azure Entra where configured.
OpenAI Compatible Gateways and providers that expose OpenAI-style endpoints. API key or no authentication, depending on provider.
Ollama Local or self-hosted models. No authentication or API key.
Anthropic Claude Claude models through Anthropic API. API key.

Mini DBA stores BYO AI settings in the browser local storage for the current user. Settings are encrypted before storage, and provider capability information may be cached so the console can adapt to the provider.

Step By Step Setup

  1. Open Settings.
  2. Select AI Provider.
  3. Turn on Use my AI provider for console AI requests.
  4. Choose the provider.
  5. Choose the authentication method.
  6. Enter a display name that makes the model obvious in the AI panel.
  7. Enter the model name or Azure deployment name.
  8. Enter the base URL or Azure endpoint where required.
  9. Enter the API key, tenant ID, managed identity client ID, or API version fields required by that provider.
  10. Save.
  11. Select Test and confirm Mini DBA can connect to the provider and model.
  12. Open a monitoring page, select Ask AI, and confirm the provider badge shows BYO AI with your display name and model.

Provider Field Guidance

For Azure OpenAI, the model field is the deployment name. The endpoint should be the Azure OpenAI resource URL, not a generic OpenAI URL. If Azure Entra authentication is used, confirm the identity has permission to call the deployment.

For OpenAI-compatible providers, use the provider's OpenAI-compatible base URL. Examples include model gateways, inference providers, or internal proxy services. Test the connection because compatibility can vary by provider and model.

For Ollama, the default base URL is commonly http://localhost:11434/v1. The browser and Mini DBA Console need network access to the endpoint you enter. If Mini DBA runs on a server and the browser runs elsewhere, localhost may not point where you expect.

For Anthropic, use the Anthropic base URL and the API version required by your account or deployment.

Token And Cost Behavior

When BYO AI is enabled, Mini DBA uses your configured provider and does not apply the default free daily Mini DBA token quota. Your provider may still charge for input tokens, output tokens, requests, or model runtime. The AI Assistant may show duration, token count, and model information when the provider returns it.

Control cost by:

  • Asking focused questions.
  • Turning off Alerts when alert context is not relevant.
  • Turning off Sessions when you do not need live session rows.
  • Using execution plan AI only when the plan is relevant.
  • Clearing the conversation when old context is no longer useful.

Troubleshooting BYO AI

If the Test action fails, check the provider name, endpoint URL, model or deployment name, API key, authentication mode, tenant ID, managed identity client ID, API version, network path, proxy rules, and provider quota. Check Web Log for console-side errors and Service Log when engine-side network or settings issues are suspected.

If the AI panel still shows Default AI, return to AI Provider settings and confirm the BYO AI provider is enabled and saved. The provider badge updates when settings are changed.

Related Pages