Deploying on Google Cloud Platform (GCP)¶

To deploy a filtering node on the Google Cloud Platform, perform the following steps:

Log in to your Google Cloud Platform account.
Launch a filtering node instance.
Configure the filtering node instance.
Connect to the filtering node instance via SSH.
Connect the filtering node to the Wallarm Cloud.
Set up the filtering node for using a proxy server.
Set up filtering and proxying rules
Allocate more memory for the Wallarm node.
Configure logging.
Restart NGINX.

1. Log in to your Google Cloud Platform account¶

2. Launch a filtering node instance¶

Launch the instance via the Google Cloud UI¶

To launch the filtering node instance via the Google Cloud UI, please open the Wallarm node image on the Google Cloud Marketplace and click LAUNCH.

The instance will launch with a preinstalled filtering node. To see detailed information on launching instances in the Google Cloud, please proceed to the official Google Cloud Platform documentation.

Launch the instance via Terraform or other tools¶

When using a tool like Terraform to launch the filtering node instance using Wallarm GCP image, you may need to provide the name of this image in the Terraform configuration.

Image name has the following format:

wallarm-node-195710/wallarm-node-<IMAGE_VERSION>-build

To launch the instance with the filtering node version 4.4, please use the following image name:
```
wallarm-node-195710/wallarm-node-4-4-20221122-092419
```

To get the image name, you can also follow these steps:

Install Google Cloud SDK.

Execute the command gcloud compute images list with the following parameters:

gcloud compute images list --project wallarm-node-195710 --filter="name~'wallarm-node-4-4-*'" --no-standard-images

Copy the version value from the name of the latest available image and paste the copied value into the provided image name format. For example, the filtering node version 4.4 image will have the following name:
```
wallarm-node-195710/wallarm-node-4-4-20221122-092419
```

3. Configure the filtering node instance¶

Perform the following actions to configure the launched filtering node instance:

Navigate to the VM instances page in the Compute Engine section of the menu.
Select the launched filtering node instance and click the Edit button.
Allow the required types of incoming traffic by ticking the corresponding checkboxes in the Firewalls setting.
If necessary, you can restrict connecting to the instance with the project SSH keys and use a custom SSH key pair for connecting to this instance. To do this, perform the following actions:
1. Tick the Block project-wide checkbox in the SSH Keys setting.
2. Click the Show and edit button in the SSH Keys setting to expand the field for entering an SSH key.
3. Generate a pair of public and private SSH keys. For example, you can use the ssh-keygen and PuTTYgen utilities.
4. Copy an open key in OpenSSH format from the interface of the used key generator (in the current example, the generated public key should be copied from the Public key for pasting into OpenSSH authorized_keys file area of the PuTTYgen interface) and paste it into the field containing the Enter entire key data hint.
5. Save the private key. It will be required for connecting to the configured instance in the future.
Click the Save button at the bottom of the page to apply the changes.

4. Connect to the filtering node instance via SSH¶

To see detailed information about ways of connecting to instances, proceed to this link.

Connecting to the instance via a custom private key

If during base instance creation process you have enabled connection to the instance via a custom SSH key pair, make sure you have access to the private key from this key pair.

5. Connect the filtering node to the Wallarm Cloud¶

The filtering node interacts with the Wallarm Cloud. To connect the node to the Cloud:

Make sure that your Wallarm account has the Administrator role enabled in Wallarm Console.

You can check mentioned settings by navigating to the users list in the US Cloud or EU Cloud.
Open Wallarm Console → Nodes in the US Cloud or EU Cloud and create the node of the Wallarm node type.
Copy the generated token.
Run the register-node script in a system with the filtering node:
US CloudEU Cloud
sudo /usr/share/wallarm-common/register-node -t <NODE_TOKEN> -H us1.api.wallarm.com
sudo /usr/share/wallarm-common/register-node -t <NODE_TOKEN>
<NODE_TOKEN> is the copied token value.

Using one token for several installations

You can use one token in several installations regardless of the selected platform. It allows logical grouping of node instances in the Wallarm Console UI. Example: you deploy several Wallarm nodes to a development environment, each node is on its own machine owned by a certain developer.

6. Set up the filtering node for using a proxy server¶

Info

This setup step is intended for users who use their own proxy server for the operation of the protected web applications.

If you do not use a proxy server, skip this step of the setup.

You need to assign new values to the environment variables, which define the proxy server used, to configure Wallarm node for using your proxy server.

Add new values of the environment variables to the /etc/environment file:

Add https_proxy to define a proxy for the https protocol.
Add http_proxy to define a proxy for the http protocol.
Add no_proxy to define the list of the resources proxy should not be used for.

Assign the <scheme>://<proxy_user>:<proxy_pass>@<host>:<port> string values to the https_proxy and http_proxy variables.

<scheme> defines the protocol used. It should match the protocol that the current environment variable sets up proxy for.
<proxy_user> defines the username for proxy authorization.
<proxy_pass> defines the password for proxy authorization.
<host> defines a host of the proxy server.
<port> defines a port of the proxy server.

Assign a "<res_1>, <res_2>, <res_3>, <res_4>, ..." array value, where <res_1>, <res_2>, <res_3>, and <res_4> are the IP addresses and/or domains, to the no_proxy variable to define a list of the resources which proxy should not be used for. This array should consist of IP addresses and/or domains.

Resources that need to be addressed without a proxy

Add the following IP addresses and domain to the list of the resources that have to be addressed without a proxy for the system to operate correctly: 127.0.0.1, 127.0.0.8, 127.0.0.9, and localhost.
The 127.0.0.8 and 127.0.0.9 IP addresses are used for the operation of the Wallarm filtering node.

The example of the correct /etc/environment file contents below demonstrates the following configuration:

HTTPS and HTTP requests are proxied to the 1.2.3.4 host with the 1234 port, using the admin username and the 01234 password for authorization on the proxy server.
Proxying is disabled for the requests sent to 127.0.0.1, 127.0.0.8, 127.0.0.9, and localhost.

https_proxy=http://admin:01234@1.2.3.4:1234
http_proxy=http://admin:01234@1.2.3.4:1234
no_proxy="127.0.0.1, 127.0.0.8, 127.0.0.9, localhost"

7. Set up filtering and proxying rules¶

The following files contain NGINX and filtering node settings:

/etc/nginx/nginx.conf defines the configuration of NGINX
/etc/nginx/conf.d/wallarm.conf defines the global configuration of Wallarm filtering node
/etc/nginx/conf.d/wallarm-status.conf defines the filtering node monitoring service configuration

You can create your own configuration files to define the operation of NGINX and Wallarm. It is recommended to create a separate configuration file with the server block for each group of the domains that should be processed in the same way.

To see detailed information about working with NGINX configuration files, proceed to the official NGINX documentation.

Wallarm directives define the operation logic of the Wallarm filtering node. To see the list of Wallarm directives available, proceed to the Wallarm configuration options page.

Configuration file example

Let us suppose that you need to configure the server to work in the following conditions:

Only HTTP traffic is processed. There are no HTTPS requests processed.
The following domains receive the requests: example.com and www.example.com.
All requests must be passed to the server 10.80.0.5.
All incoming requests are considered less than 1MB in size (default setting).
The processing of a request takes no more than 60 seconds (default setting).
Wallarm must operate in the monitor mode.
Clients access the filtering node directly, without an intermediate HTTP load balancer.

Creating a configuration file

You can create a custom NGINX configuration file (e.g. example.com.conf) or modify the default NGINX configuration file (default.conf).

When creating a custom configuration file, make sure that NGINX listens to the incoming connections on the free port.

To meet the listed conditions, the contents of the configuration file must be the following:

    server {
      listen 80;
      listen [::]:80 ipv6only=on;

      # the domains for which traffic is processed
      server_name example.com; 
      server_name www.example.com;

      # turn on the monitoring mode of traffic processing
      wallarm_mode monitoring; 
      # wallarm_application 1;

      location / {
        # setting the address for request forwarding
        proxy_pass http://10.80.0.5; 
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
      }
    }

8. Allocate more memory for the Wallarm node¶

The Wallarm node uses Tarantool, an open‑source in-memory database, to calculate traffic metrics required for automated adjusting of security rules.

By default, the amount of RAM allocated to Tarantool is 40% of the total instance memory.

You can change the amount of RAM allocated for Tarantool. To allocate the instance RAM to Tarantool:

Open the Tarantool configuration file:

sudo vim /etc/default/wallarm-tarantool

Set the amount of allocated RAM in the SLAB_ALLOC_ARENA in GB. The value can be an integer or a float (a dot . is a decimal separator).

For production environments, the recommended amount of RAM allocated for the postanalytics module is 75% of the total server memory. If testing the Wallarm node or having a small instance size, the lower amount can be enough (e.g. 25% of the total memory).

For example:
If testing the nodeIf deploying the node to the production environment
SLAB_ALLOC_ARENA=0.5
SLAB_ALLOC_ARENA=24
To apply changes, restart the Tarantool daemon:
```
sudo systemctl restart wallarm-tarantool
```

9. Configure logging¶

Configure the filtering node variables logging using NGINX. This will allow to perform a quick filtering node diagnostics with the help of the NGINX log file.

10. Restart NGINX¶

Restart NGINX by running the following command:

sudo systemctl restart nginx

The installation is complete¶

The installation is now complete.

Check that the filtering node runs and filters the traffic. See Check the filtering node operation.

Default settings

A freshly installed filtering node operates in blocking mode (see the wallarm_mode directive description) by default.

This may result in the inoperable Wallarm Scanner. If you plan to use the Scanner, then you need to perform additional actions to render Scanner operational.

Additional settings¶

The filtering node may require some additional configuration after installation.

The document below lists a few of the typical setups that you can apply if needed.

To get more information about other available settings, proceed to the Configuration section of the Administrator’s guide.

Configuring the display of the client's real IP¶

If the filtering node is deployed behind a proxy server or load balancer without any additional configuration, the request source address may not be equal to the actual IP address of the client. Instead, it may be equal to one of the IP addresses of the proxy server or the load balancer.

In this case, if you want the filtering node to receive the client's IP address as a request source address, you need to perform an additional configuration of the proxy server or the load balancer.

Limiting the single request processing time¶

Use the wallarm_process_time_limit Wallarm directive to specify the limit of the duration for processing a single request by the filtering node.

If processing the request consumes more time than specified in the directive, then the information on the error is entered into the log file and the request is marked as an overlimit_res attack.

Limiting the server reply waiting time¶

Use the proxy_read_timeout NGINX directive to specify the timeout for reading the proxy server reply.

If the server sends nothing during this time, the connection is closed.

Limiting the maximum request size¶

Use the client_max_body_size NGINX directive to specify the limit for the maximum size of the body of the client's request.

If this limit is exceeded, NGINX replies to the client with the 413 (Payload Too Large) code, also known as the Request Entity Too Large message.