update

Author: cloga

articles/machine-learning/prompt-flow/how-to-secure-prompt-flow.md

@@ -87,7 +87,7 @@ Workspace managed virtual network is the recommended way to support network isol
 - If you have strict outbound rule, make sure you have open the [Required public internet access](../how-to-secure-workspace-vnet.md#required-public-internet-access).
 - Add workspace MSI as `Storage File Data Privileged Contributor` to storage account linked with workspace. Please follow step 2 in [Secure prompt flow with workspace managed virtual network](#secure-prompt-flow-with-workspace-managed-virtual-network).
 - Meanwhile, you can follow [private Azure Cognitive Services](../../ai-services/cognitive-services-virtual-networks.md) to make them as private.
-- If you want to deploy prompt flow in workspace which secured by your own virtual network, you can deploy it to AKS cluster which is in the same virtual network. You can follow [Secure Azure Kubernetes Service inferencing environment](../how-to-secure-kubernetes-inferencing-environment.md) to secure your AKS cluster.
+- If you want to deploy prompt flow in workspace which secured by your own virtual network, you can deploy it to AKS cluster which is in the same virtual network. You can follow [Secure Azure Kubernetes Service inferencing environment](../how-to-secure-kubernetes-inferencing-environment.md) to secure your AKS cluster. Learn more about [How to deploy prompt flow to ASK cluster via code](./how-to-deploy-to-code) to secure your RAG workflows.
 - You can either create private endpoint to the same virtual network or leverage virtual network peering to make them communicate with each other.
 
 ## Known limitations

articles/machine-learning/prompt-flow/tools-reference/llm-tool.md

@@ -25,6 +25,8 @@ Prompt flow provides a few different large language model APIs:
 
 > [!NOTE]
 > We removed the `embedding` option from the LLM tool API setting. You can use an embedding API with the [embedding tool](embedding-tool.md).
+> Only key-based authentication is supported for Azure OpenAI connection.
+> Please don't use non-ascii characters in resource group name of Azure OpenAI resource, prompt flow didn't support this case.
 
 ## Prerequisites
 

articles/machine-learning/prompt-flow/tools-reference/troubleshoot-guidance.md

@@ -144,9 +144,9 @@ Check if this compute instance is assigned to you and you have access to the wor
 
 This error occurs because you're cloning a flow from others that's using a compute instance as the runtime. Because the compute instance runtime is user isolated, you need to create your own compute instance runtime or select a managed online deployment/endpoint runtime, which can be shared with others.
 
-### Find Python packages installed in runtime
+### Find Python packages installed in CI runtime
 
-Follow these steps to find Python packages installed in runtime:
+Follow these steps to find Python packages installed in CI runtime:
 
 - Add a Python node in your flow.
 - Put the following code in the code section:
@@ -165,3 +165,18 @@ Follow these steps to find Python packages installed in runtime:
 - Run the flow. Then you can find `packages.txt` in the flow folder.
 
   :::image type="content" source="../media/faq/list-packages.png" alt-text="Screenshot that shows finding Python packages installed in runtime." lightbox = "../media/faq/list-packages.png":::
+
+### Runtime start failures using custom environment
+
+#### CI (Compute instance) runtime start failure using custom environment
+
+To use promptflow as runtime on CI, you need use the base image provide by promptflow. If you want to add extra packages to the base image, you need follow the [Customize environment with Docker context for runtime](../how-to-customize-environment-runtime.md) to create a new environment. Then use it to create CI runtime.
+
+If you got `UserError: FlowRuntime on compute instance is not ready`, you need login into to terminal of CI and run `journalctl -u c3-progenitor.serivice` to check the logs.
+
+#### Automatic runtime start failure with requirements.txt or custom base image
+
+Automatic runtime support to use `requirements.txt` or custom base image in `flow.dag.yaml` to customize the image. We would recommend you to use `requirements.txt` for common case, which will use `pip install -r requirements.txt` to install the packages. If you have dependency more then python packages, you need follow the [Customize environment with Docker context for runtime](../how-to-customize-environment-runtime.md) to create build a new image base on top of promptflow base image. Then use it in `flow.dag.yaml`. Learn more about [Customize environment with Docker context for runtime](../how-to-create-manage-runtime.md#update-an-automatic-runtime-preview-on-a-flow-page).
+
+- You can not use arbitrary base image to create runtime, you need use the base image provide by promptflow.
+- Don't pin the version of `promptflow` and `promptflow-tools` in `requirements.txt`, because we already include them in the runtime base image. Using old version of `promptflow` and `promptflow-tools` may cause unexpected behavior.