scenarios/workload-genai/policies/genai-policy.xml

<policies>
    <inbound>
        <base />
        
        <!-- sets the backed to the load balanced pool-->
        <include-fragment fragment-id="simple-priority-weighted" />

        <!-- Rate limiting: applies tokens constraint defined as per the fragment, doesn't set the backend-->
        <!-- there are other variants that can be used as well-->
        <include-fragment fragment-id="rate-limiting-by-tokens" />

        <!-- tracks token consumption as custom metrics in App insights -->
        <include-fragment fragment-id="usage-tracking-with-appinsights" />
    </inbound>
    <backend>
        <!-- either base or the fragment needs to be present.  -->
        <!-- <base /> -->

        <!-- set the count to the number of backends-->
        <retry condition="@(context.Response.StatusCode == 429)" count="3" interval="1" first-fast-retry="true">
            <forward-request buffer-request-body="true" />
        </retry>
        <!-- END: sets backend service -->
    </backend>
    <outbound>
        <base />
        <!-- tracks token consumption using the event hub, this can be used as a workaround. -->
        <!-- <include-fragment fragment-id="usage-tracking-with-eventhub" /> -->
    </outbound>
    <on-error>
        <base />
    </on-error>
</policies>