diff --git a/managed_vms/cloudsql/README.md b/managed_vms/cloudsql/README.md
index b6e09b31a039..2ddefe8634bf 100644
--- a/managed_vms/cloudsql/README.md
+++ b/managed_vms/cloudsql/README.md
@@ -6,24 +6,47 @@ This sample demonstrates how to use [Google Cloud SQL](https://cloud.google.com/
 
 Before you can run or deploy the sample, you will need to do the following:
 
-1. Create a Cloud SQL instance. You can do this from the [Google Developers Console](https://console.developers.google.com) or via the [Cloud SDK](https://cloud.google.com/sdk). To create it via the SDK use the following command:
+1. Create a [Second Generation Cloud SQL](https://cloud.google.com/sql/docs/create-instance) instance. You can do this from the [Cloud Console](https://console.developers.google.com) or via the [Cloud SDK](https://cloud.google.com/sdk). To create it via the SDK use the following command:
 
-        $ gcloud sql instances create [your-instance-name] \
-            --assign-ip \
-            --authorized-networks 0.0.0.0/0 \
-            --tier D0
+        $ gcloud sql instances create YOUR_INSTANCE_NAME \
+            --activation-policy=ALWAYS \
+            --tier=db-n1-standard-1
 
-2. Create a new user and database for the application. The easiest way to do this is via the [Google Developers Console](https://console.developers.google.com/project/_/sql/instances/example-instance2/access-control/users). Alternatively, you can use MySQL tools such as the command line client or workbench, but you will need to set a root password for your database using `gcloud sql instances set-root-password`.
+1. Set the root password on your Cloud SQL instance:
 
-3. Update the connection string in ``app.yaml`` with your instance values.
+        $ gcloud sql instances set-root-password YOUR_INSTANCE_NAME --password YOUR_INSTANCE_ROOT_PASSWORD
 
-4. Finally, run ``create_tables.py`` to ensure that the database is properly configured and to create the tables needed for the sample.
+1. Create a [Service Account](https://cloud.google.com/sql/docs/external#createServiceAccount) for your project. You'll use this service account to connect to your Cloud SQL instance locally.
+
+1. Download the [Cloud SQL Proxy](https://cloud.google.com/sql/docs/sql-proxy).
+
+1. Run the proxy to allow connecting to your instance from your machine.
+
+    $ cloud_sql_proxy \
+        -dir /tmp/cloudsql \
+        -instances=YOUR_PROJECT_ID:us-central1:YOUR_INSTANCE_NAME=tcp:3306 \
+        -credential_file=PATH_TO_YOUR_SERVICE_ACCOUNT_JSON
+
+1. Use the MySQL command line tools (or a management tool of your choice) to create a [new user](https://cloud.google.com/sql/docs/create-user) and [database](https://cloud.google.com/sql/docs/create-database) for your application:
+
+    $ mysql -h 127.0.0.1 -u root -p
+    mysql> create database YOUR_DATABASE;
+    mysql> create user 'YOUR_USER'@'%' identified by 'PASSWORD';
+    mysql> grant all on YOUR_DATABASE.* to 'YOUR_USER'@'%';
+
+1. Set the connection string environment variable. This allows the app to connect to your Cloud SQL instance through the proxy:
+
+    export SQLALCHEMY_DATABASE_URI=mysql+pymysql://USER:PASSWORD@127.0.0.1/YOUR_DATABASE
+
+1. Run ``create_tables.py`` to ensure that the database is properly configured and to create the tables needed for the sample.
+
+1. Update the connection string in ``app.yaml`` with your configuration values. These values are used when the application is deployed.
 
 ## Running locally
 
 Refer to the [top-level README](../README.md) for instructions on running and deploying.
 
-You will need to set the following environment variables via your shell before running the sample:
+It's recommended to follow the instructions above to run the Cloud SQL proxy. You will need to set the following environment variables via your shell before running the sample:
 
     $ export SQLALCHEMY_DATABASE_URI=[your connection string]
     $ python main.py
diff --git a/managed_vms/cloudsql/app.yaml b/managed_vms/cloudsql/app.yaml
index c2b0c5066649..ea5e13a4078a 100644
--- a/managed_vms/cloudsql/app.yaml
+++ b/managed_vms/cloudsql/app.yaml
@@ -7,11 +7,15 @@ runtime_config:
 
 #[START env]
 env_variables:
-    # Replace user, password, and host with the values obtained when
-    # configuring your Cloud SQL instance.
-    SQLALCHEMY_DATABASE_URI: mysql+pymysql://user:password@host/db
-    # If you are connecting over SSL, you can specify your certificates by
-    # using a connection string such as:
-    #    > mysql+pymysql://user:password@host/db?
-    #    ssl_key=client-key.pem?ssl_cert=client-cert.pem?ssl_ca=server-ca.pem
+    # Replace user, password, database, project, and instance with the values obtained
+    # when configuring your Cloud SQL instance.
+    SQLALCHEMY_DATABASE_URI: >-
+      mysql+pymysql://USER:PASSWORD@/DATABASE?unix_socket=/cloudsql/PROJECT:us-central1:INSTANCE
 #[END env]
+
+#[START cloudsql_settings]
+# Replace project and instance with the values obtained  when configuring your
+# Cloud SQL instance.
+beta_settings:
+    cloud_sql_instances: PROJECT:us-central1:INSTANCE
+#[END cloudslq_settings]
diff --git a/managed_vms/cloudsql/main.py b/managed_vms/cloudsql/main.py
index 1b22c31ad491..baff9df43f9e 100644
--- a/managed_vms/cloudsql/main.py
+++ b/managed_vms/cloudsql/main.py
@@ -35,6 +35,7 @@ def is_ipv6(addr):
 # [START example]
 # Environment variables are defined in app.yaml.
 app.config['SQLALCHEMY_DATABASE_URI'] = os.environ['SQLALCHEMY_DATABASE_URI']
+app.config['SQLALCHEMY_TRACK_MODIFICATIONS'] = False
 
 db = SQLAlchemy(app)