Skip to content

What is new in Wallarm node (if upgrading an EOL node)

This page lists the changes available when upgrading the node of the deprecated version (3.6 and lower) up to version 4.6. Listed changes are available for both the regular (client) and multi-tenant Wallarm nodes.

Wallarm nodes 3.6 and lower are deprecated

Wallarm nodes 3.6 and lower are recommended to be upgraded since they are deprecated.

Node configuration and traffic filtration have been significantly simplified in the Wallarm node of version 4.x. Some settings of node 4.x are incompatible with the nodes of older versions. Before upgrading the modules, please carefully review the list of changes and general recommendations.

Breaking changes due to the deleted metrics

Starting from version 4.0, the Wallarm node does not collect the following collectd metrics:

  • curl_json-wallarm_nginx/gauge-requests - you can use the curl_json-wallarm_nginx/gauge-abnormal metric instead

  • curl_json-wallarm_nginx/gauge-attacks

  • curl_json-wallarm_nginx/gauge-blocked

  • curl_json-wallarm_nginx/gauge-time_detect

  • curl_json-wallarm_nginx/derive-requests

  • curl_json-wallarm_nginx/derive-attacks

  • curl_json-wallarm_nginx/derive-blocked

  • curl_json-wallarm_nginx/derive-abnormal

  • curl_json-wallarm_nginx/derive-requests_lost

  • curl_json-wallarm_nginx/derive-tnt_errors

  • curl_json-wallarm_nginx/derive-api_errors

  • curl_json-wallarm_nginx/derive-segfaults

  • curl_json-wallarm_nginx/derive-memfaults

  • curl_json-wallarm_nginx/derive-softmemfaults

  • curl_json-wallarm_nginx/derive-time_detect

Rate limits

The lack of proper rate limiting has been a significant problem for API security, as attackers can launch high-volume requests causing a denial of service (DoS) or overload the system, which hurts legitimate users.

With Wallarm's rate limiting feature supported since Wallarm node 4.6, security teams can effectively manage the service's load and prevent false alarms, ensuring that the service remains available and secure for legitimate users. This feature offers various connection limits based on request and session parameters, including traditional IP-based rate limiting, JSON fields, base64 encoded data, cookies, XML fields, and more.

For example, you can limit API connections for each user, preventing them from making thousands of requests per minute. This would put a heavy load on your servers and could cause the service to crash. By implementing rate limiting, you can protect your servers from overload and ensure that all users have fair access to the API.

You can configure rate limits easily in the Wallarm Console UI → Rules → Set rate limit by specifying the rate limit scope, rate, burst, delay, and response code for your particular use case.

Guide on rate limit configuration →

Although the rate limiting rule is the recommended method for setting up the feature, you can also configure rate limits using the new NGINX directives:

Detection of the new attack types

Wallarm detects new attack types:

  • Broken Object Level Authorization (BOLA), also known as Insecure Direct Object References (or IDOR), became one of the most common API vulnerabilities. When an application includes an IDOR / BOLA vulnerability, it has a strong probability of exposing sensitive information or data to attackers. All the attackers need to do is exchange the ID of their own resource in the API call with an ID of a resource belonging to another user. The absence of proper authorization checks enables attackers to access the specified resource. Thus, every API endpoint that receives an ID of an object and performs any type of action on the object can be an attack target.

    To prevent exploitation of this vulnerability, Wallarm node 4.2 and above contain a new trigger which you can use to protect your endpoints from BOLA attacks. The trigger monitors the number of requests to a specified endpoint and creates a BOLA attack event when thresholds from the trigger are exceeded.

  • Mass Assignment

    During a Mass Assignment attack, attackers try to bind HTTP request parameters into program code variables or objects. If an API is vulnerable and allows binding, attackers may change sensitive object properties that are not intended to be exposed, which could lead to privilege escalation, bypassing security mechanisms, and more.

  • SSRF

    A successful SSRF attack may allow an attacker to make requests on behalf of the attacked web server; this potentially leads to revealing the web application's network ports in use, scanning the internal networks, and bypassing authorization.

Checking JSON Web Token strength

JSON Web Token (JWT) is a popular authentication standard used to exchange data between resources like APIs securely. JWT compromisation is a common aim of attackers as breaking authentication mechanisms provides them full access to web applications and APIs. The weaker JWTs, the higher chance for it to be compromised.

Starting from version 4.4, you can enable Wallarm to detect the following JWT weaknesses:

  • Unencrypted JWTs

  • JWTs signed using compromised secret keys

To enable, use the Weak JWT trigger.

Checking JSON Web Tokens for attacks

JSON Web Token (JWT) is one of the most popular authentication methods. This makes it a favorite tool to perform attacks (for example SQLi or RCE) that are very difficult to find because the data in the JWT is encoded and it can be located anywhere in the request.

Wallarm node 4.2 and above find the JWT anywhere in the request, decodes it and blocks (in the appropriate filtration mode) any attack attempts through this authentication method.

Supported installation options

  • Wallarm Ingress controller based on the latest version of Community Ingress NGINX Controller, 1.7.1.

    Instructions on migrating to the Wallarm Ingress controller →

  • Added support for AlmaLinux, Rocky Linux and Oracle Linux 8.x instead of the deprecated CentOS 8.x.

    Wallarm node packages for the alternative operating systems will be stored in the CentOS 8.x repository.

  • Added support for Debian 11 Bullseye

  • Added support for Ubuntu 22.04 LTS (jammy)

  • Dropped support for CentOS 6.x (CloudLinux 6.x)

  • Dropped support for Debian 9.x

  • Dropped support for Debian 10.x for Wallarm to be installed as the module for either NGINX stable or NGINX Plus

  • Dropped support for the operating system Ubuntu 16.04 LTS (xenial)

  • Version of Envoy used in Wallarm Envoy-based Docker image has been increased to 1.18.4

See the full list of supported installation options →

New method for the serverless Wallarm node deployment

The new deployment method lets you configure the Wallarm CDN node outside your infrastructure in 15 minutes. You need to just point to the domain to be protected and add the Wallarm CNAME record to the domain's DNS records.

Instructions on the CDN node deployment

System requirements for the filtering node installation

  • The filtering node now supports IP address allowlisting, denylisting, and graylisting. Wallarm Console allows adding both single IPs and countries or data centers to any IP list type.

    The Wallarm node downloads an actual list of IP addresses registered in allowlisted, denylisted, or graylisted countries, regions or data centers from GCP storage. By default, access to this storage can be restricted in your system. Allowing access to GCP storage is a new requirement for the virtual machine to install the filtering node.

    Range of GCP IP addresses that should be allowed →

  • The filtering node now uploads data to the Cloud using us1.api.wallarm.com:443 (US Cloud) and api.wallarm.com:443 (EU Cloud) instead of us1.api.wallarm.com:444 and api.wallarm.com:444.

    If your server with the deployed node has a limited access to the external resources and the access is granted to each resource separately, after upgrade to version 4.x the synchronization between the filtering node and the Cloud will stop. The upgraded node needs to be granted access to the API endpoint with the new port.

Unified registration of nodes in the Wallarm Cloud by tokens

With the release of Wallarm node 4.6, email-password based registration of Wallarm nodes in the Cloud has been removed. It is now mandatory to switch to the new token-based node registration method to continue with Wallarm node 4.6.

The release 4.6 enables you to register the Wallarm node in the Wallarm Cloud by the token on any supported platform, which ensures a more secure and faster connection to the Wallarm Cloud as follows:

  • Dedicated user accounts of the Deploy role allowing only to install the node are no longer required.

  • Users' data remains securely stored in the Wallarm Cloud.

  • Two-factor authentication enabled for the user accounts does not prevent nodes from being registered in the Wallarm Cloud.

  • The initial traffic processing and request postanalytics modules deployed to separate servers can be registered in the Cloud by one node token.

Changes in node registration methods result in some updates in node types:

  • The node supporting the unified registration by token has the Wallarm node type. The script to be run on the server to register the node is named register-node.

    Previously, the Wallarm node was named cloud node. It also supported registration by the token but with the different script named addcloudnode.

    The cloud node is not required to be migrated to the new node type.

  • The regular node supporting the registration by "email-password" passed to the addnode script is deprecated.

    Starting from version 4.0, registration of the node deployed as the NGINX, NGINX Plus module or the Docker container looks as follows:

    1. Create the Wallarm node in Wallarm Console and copy the generated token.
    2. Run the register-node script with the node token passed or run the Docker container with the WALLARM_API_TOKEN variable defined.

    Regular node support

    The regular node type is deprecated in release 4.x and will be removed in future releases.

    It is recommended to replace the regular node with the Wallarm node before the regular type is removed. You will find the appropriate instructions in the node upgrade guides.

Terraform module to deploy Wallarm on AWS

Starting from release 4.0, you can easily deploy Wallarm to AWS from the Infrastructure as Code (IaC)-based environment using the Wallarm Terraform module.

The Wallarm Terraform module is the scalable solution meeting the best industry standards of security and failover ensuring. During its deployment, you can choose either the proxy or mirror deployment option based on your requirements for the traffic flow.

We have also prepared the usage examples for both deployment options involving basic deployment configurations as well as advanced ones compatible with such solutions as AWS VPC Traffic Mirroring.

Documentation on the Wallarm Terraform module for AWS

Wallarm AWS image distributed with the ready-to-use cloud-init.py script

If following the Infrastructure as Code (IaC) approach, you may need to use the cloud-init script to deploy the Wallarm node to AWS. Starting from release 4.0, Wallarm distributes its AWS cloud image with the ready‑to‑use cloud-init.py script.

Specification of the Wallarm cloud-init script

Simplified multi-tenant node configuration

For the multi-tenant nodes, tenants and applications are now defined each with its own directive:

Instructions on the multi-tenant node upgrade

Filtration modes

  • New safe blocking filtration mode.

    This mode enables a significant reduction of false positive number by blocking only malicious requests originating from graylisted IP addresses.

  • Analysis of request sources is now performed only in the safe_blocking and block modes.

    • If the Wallarm node operating in the off or monitoring mode detects the request originating from the denylisted IP, it does not block this request.
    • Wallarm node operating in the monitoring mode uploads all the attacks originating from the allowlisted IP addresses to the Wallarm Cloud.

More details on Wallarm node modes →

Request source control

The following parameters for request source control have been deprecated:

There are the following new features for request source control:

  • Wallarm Console section for full IP address allowlist, denylist and graylist control.

  • Support for new filtration mode safe_blocking and IP address graylists.

    The safe blocking mode enables a significant reduction of false positive number by blocking only malicious requests originating from graylisted IP addresses.

    For automatic IP address graylisting there is a new trigger Add to graylist released.

  • Automated allowlisting of Wallarm Vulnerability Scanner IP addresses. Manual allowlisting of Scanner IP addresses is no longer required.

  • Ability to allowlist, denylist, or graylist a subnet, Tor network IPs, VPN IPs, a group of IP addresses registered in a specific country, region, or data center.

  • Ability to allowlist, denylist, or graylist request sources for specific applications.

  • New NGINX directive and Envoy parameter disable_acl to disable request origin analysis.

    Details on the disable_acl NGINX directive →

    Details on the disable_acl Envoy parameter →

Details on adding IPs to the allowlist, denylist, and graylist →

New module for API inventory discovery

New Wallarm nodes are distributed with the module API Discovery automatically identifying the application API. The module is disabled by default.

Details on the API Discovery module →

Enhanced attack analysis with the libdetection library

Attack analysis performed by Wallarm has been enhanced by involving an additional attack validation layer. Wallarm node 4.4 and above in all form-factors (including Envoy) are distributed with the libdetection library enabled by default. This library performs secondary fully grammar-based validation of all SQLi attacks reducing the number of false positives detected among SQL injections.

Memory consumption increase

With the libdetection library enabled, the amount of memory consumed by NGINX/Envoy and Wallarm processes may increase by about 10%.

Details on how Wallarm detects attacks →

The rule enabling the overlimit_res attack detection fine-tuning

We have introduced the new rule allowing the overlimit_res attack detection fine-tuning.

The overlimit_res attack detection fine-tuning via the NGINX and Envoy configuration files is considered to be the deprecated way:

  • The rule allows setting up a single request processing time limit as the wallarm_process_time_limit NGINX directive and process_time_limit Envoy parameter did before.

  • The rule allows to block or pass the overlimit_res attacks in accordance with the node filtration mode instead of the wallarm_process_time_limit_block NGINX directive and process_time_limit_block Envoy parameter configuration.

The listed directives and parameters have been deprecated and will be deleted in future releases. It is recommended to transfer the overlimit_res attack detection configuration from directives to the rule before. Relevant instructions are provided for each node deployment option.

If the listed parameters are explicitly specified in the configuration files and the rule is not created yet, the node processes requests as set in the configuration files.

New blocking page

The sample blocking page /usr/share/nginx/html/wallarm_blocked.html has been updated. In the new node version, it has new layout and supports the logo and support email customization.

New blocking page with the new layout looks as follows by default:

Wallarm blocking page

More details on the blocking page setup →

New parameters for basic node setup

Disabling IPv6 connections for the NGINX-based Wallarm Docker container

The NGINX-based Wallarm Docker image 4.2 and above supports the new environment variable DISABLE_IPV6. This variable enables you to prevent NGINX from IPv6 connection processing, so that it only can process IPv4 connections.

Renamed parameters, files and metrics

  • The following NGINX directives and Envoy parameters have been renamed:

    Parameters with previous names are still supported but will be deprecated in future releases. The parameter logic has not changed.

  • The Ingress annotation nginx.ingress.kubernetes.io/wallarm-instance has been renamed to nginx.ingress.kubernetes.io/wallarm-application.

    The annotation with the previous name is still supported but will be deprecated in future releases. The annotation logic has not changed.

  • The file with the custom ruleset build /etc/wallarm/lom has been renamed to /etc/wallarm/custom_ruleset. In the file system of new node versions, there is only the file with the new name.

    Default values of the NGINX directive wallarm_custom_ruleset_path and Envoy parameter custom_ruleset have been changed appropriately. New default value is /etc/wallarm/custom_ruleset.

  • The private key file /etc/wallarm/license.key has been renamed to /etc/wallarm/private.key. Starting from the node version 4.0 the new name is used by default.

  • The collectd metric gauge-lom_id has been renamed to gauge-custom_ruleset_id.

    In new node versions, the collectd service collects both the deprecated and new metrics. The deprecated metric collection will be stopped in future releases.

    All collectd metrics →

  • The /var/log/wallarm/addnode_loop.log log file in the Docker containers has been renamed to /var/log/wallarm/registernode_loop.log.

Parameters of the statistics service

  • The Wallarm statistics service returns the new rate_limit parameters with the Wallarm rate limiting module data. New parameters cover rejected and delayed requests, as well as indicate any problems with the module's operation.

  • The number of requests originating from denylisted IPs is now displayed in the statistic service output, in the new parameter blocked_by_acl and in the existing parameters requests, blocked.

  • The service return one more new parameter custom_ruleset_ver which points to the custom ruleset format being used by Wallarm nodes.

  • The following node statistics parameters have been renamed:

    • lom_apply_time → custom_ruleset_apply_time
    • lom_id → custom_ruleset_id

    In new node versions, the http://127.0.0.8/wallarm-status endpoint temporarily returns both the deprecated and new parameters. The deprecated parameters will be removed from the service output in future releases.

Details on the statistics service →

New variables to configure the node logging format

The following node logging variables have been changed:

  • wallarm_request_time has been renamed to wallarm_request_cpu_time

    This variable means time in seconds the CPU spent processing the request.

    The variable with the previous name is deprecated and will be removed in future releases. The variable logic has not changed.

  • wallarm_request_mono_time has been added

    This variable means time in seconds the CPU spent processing the request + time in the queue.

Increasing the performance by omitting attack search in requests from denylisted IPs

The new wallarm_acl_access_phase directive enables you to increase the Wallarm node performance by omitting the attack search stage during the analysis of requests from denylisted IPs. This configuration option is useful if there are many denylisted IPs (e.g. the whole countries) producing high traffic that heavily loads the working machine CPU.

Easy grouping for node instances

Now you can easily group node instances using one API token with the Deploy role for their installation together with the WALLARM_LABELS variable and its group label.

For example: 

docker run -d -e WALLARM_API_TOKEN='<API TOKEN WITH DEPLOY ROLE>' -e NGINX_BACKEND='example.com' -e WALLARM_API_HOST='us1.api.wallarm.com' -e WALLARM_LABELS='group=<GROUP>' -p 80:80 wallarm/node:4.6.1-1

...will place node instance into the <GROUP> instance group (existing, or, if does not exist, it will be created).

Upgrade process

  1. Review recommendations for the modules upgrade.

  2. Upgrade installed modules following the instructions for your Wallarm node deployment option:

  3. Migrate allowlist and denylist configuration from previous Wallarm node versions to 4.6.

Other updates in Wallarm products and components →