update to wait for deployments

Author: madiepev

Labs/11/Deploy to batch endpoint.ipynb

@@ -218,6 +218,7 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "nteract": {
@@ -227,6 +228,9 @@
         }
       },
       "source": [
+        "> **IMPORTANT!**\n",
+        "> Wait until the endpoint is created before continuing! A green notification should appear in the studio. \n",
+        "\n",
         "## Create the deployment\n",
         "\n",
         "A deployment is a set of resources required for hosting the model that does the actual inferencing. We will create a deployment for our endpoint using the `BatchDeployment` class. \n",
@@ -294,9 +298,13 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {},
       "source": [
+        "> **IMPORTANT!**\n",
+        "> Wait until the deployment is completed before continuing! A green notification should appear in the studio. \n",
+        "\n",
         "You can deploy multiple models to an endpoint. You can set the default deployment to specify which model should be used by default when calling a batch endpoint."
       ]
     },
@@ -330,9 +338,12 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {},
       "source": [
+        "> **IMPORTANT!** Wait until the default deployment is set before continuing! A green notification should appear in the studio. \n",
+        "\n",
         "## Prepare the data for batch predictions\n",
         "\n",
         "In the `data` folder you'll find CSV files with unlabeled data. You'll create a data asset that points to the files in the `data` folder, which you'll use as input for the batch job."

Labs/11/Deploy to online endpoint.ipynb

@@ -176,6 +176,7 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "nteract": {
@@ -185,6 +186,8 @@
         }
       },
       "source": [
+        "> **IMPORTANT!** Wait until the endpoint is created successfully before continuing! A green notification should appear in the studio. \n",
+        "\n",
         "## Configure the deployment\n",
         "\n",
         "You can deploy multiple models to an endpoint. This is mostly useful when you want to update the deployed model while keeping the current model in production. You'll need to configure the deployment to specify which model needs to be deployed to an endpoint. In the following cell, you'll refer to the model trained and stored in the local `model` folder (stored in the same folder as this notebook). Note that since you're working with an MLflow model, you don't need to specify the environment or scoring script.\n",
@@ -274,11 +277,14 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {},
       "source": [
         "The deployment of the model may take 10-15 minutes. While waiting for the model to be deployed, you can learn more about [managed endpoints in this video](https://www.youtube.com/watch?v=SxFGw_OBxNM&ab_channel=MicrosoftDeveloper).\n",
         "\n",
+        "> **IMPORTANT!** Wait until the deployment is completed before continuing! A green notification should appear in the studio. \n",
+        "\n",
         "Since you only have one model deployed to the endpoint, you want this deployment to take 100% of the traffic. If you deploy multiple models to the endpoint, you could use the same approach to distribute traffic across the deployed models."
       ]
     },
@@ -310,6 +316,7 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "nteract": {
@@ -319,6 +326,8 @@
         }
       },
       "source": [
+        "> **IMPORTANT!** Wait until the blue deployment is configured before continuing! A green notification should appear in the studio. \n",
+        "\n",
         "## Test the deployment\n",
         "\n",
         "Let's test the deployed model by invoking the endpoint. A JSON file with sample data is used as input. The trained model predicts whether a patient has diabetes or not, based on medical data like age, BMI, and the number of pregnancies. A `[0]` indicates a patient doesn't have diabetes. A `[1]` means a patient does have diabetes."