Configuration options for the NGINX‑based Wallarm node¶

Learn fine-tuning options available for the Wallarm NGINX modules to get the most out of the Wallarm solution.

Wallarm directives¶

disable_acl¶

Allows disabling analysis of requests origins. If disabled (on), the filtering node does not download IP lists from the Wallarm Cloud and skips request source IPs analysis.

wallarm_acl_access_phase¶

The directive forces the NGINX-based Wallarm node to block requests originating from denylisted IPs at the NGINX access phase which means:

• With wallarm_acl_access_phase on, the Wallarm node immediately blocks any requests from denylisted IPs in any filtration mode (except off) and does not search attack signs in requests from denylisted IPs.

  This is the default and recommended value since it makes denylists to work standardly and significantly reduces the load of the CPU of the node.

• With wallarm_acl_access_phase off, the Wallarm node analyzes requests for attack signs first and then if operating in the block or safe_blocking mode blocks requests originating from denylisted IPs.

  In the monitoring filtration mode, the node searches for attack signs in all requests but never block them even if the source IP is denylisted.

  The Wallarm node behavior with wallarm_acl_access_phase off significantly increases the load of the CPU of the node.

wallarm_acl_export_enable¶

The directive enables on / disables off sending statistics about the requests from the denylisted IPs from node to the Cloud.

• With wallarm_acl_export_enable on the statistics on the requests from the denylisted IPs will be displayed in the Attacks section.

• With wallarm_acl_export_enable off the statistics on the requests from the denylisted IPs will not be displayed.

wallarm_api_conf¶

A path to the node.yaml file, which contains access requirements for the Wallarm API.

Example:

wallarm_api_conf /etc/wallarm/node.yaml

# Docker NGINX-based image, cloud image and all-in-one installer installations
# wallarm_api_conf /opt/wallarm/etc/wallarm/node.yaml

Used to upload serialized requests from the filtering node directly to the Wallarm API (Cloud) instead of uploading into the postanalytics module (Tarantool).

Only requests with attacks are sent to the API. Requests without attacks are not saved.

Example of the node.yaml file content:

# API connection credentials

hostname: <some name>
uuid: <some uuid>
secret: <some secret>

# API connection parameters (the parameters below are used by default)

host: api.wallarm.com
port: 443
ca_verify: true

wallarm_application¶

Unique identifier of the protected application to be used in the Wallarm Cloud. The value can be a positive integer except for 0.

Unique identifiers can be set for both the application domains and the domain paths, for example:

server {
    listen 80 default_server;
    listen [::]:80 default_server ipv6only=on;
    listen 443 ssl;

    ...

    wallarm_mode monitoring;
    wallarm_application 1;
    location / {
            proxy_pass http://example.com;
            include proxy_params;
    }
}

server {
    listen 80 default_server;
    listen [::]:80 default_server ipv6only=on;
    listen 443 ssl;

    ...

    wallarm_mode monitoring;
    wallarm_application 2;
    location / {
            proxy_pass http://test.com;
            include proxy_params;
    }
}

server {
    listen 80 default_server;
    listen [::]:80 default_server ipv6only=on;
    listen 443 ssl;

    ...

    wallarm_mode monitoring;
    location /login {
            proxy_pass http://example.com/login;
            include proxy_params;
            wallarm_application 3;
    }

    location /users {
            proxy_pass http://example.com/users;
            include proxy_params;
            wallarm_application 4;
    }
}

More details on setting up applications →

wallarm_block_page¶

Lets you set up the response to the blocked request.

More details on the blocking page and error code configuration →

wallarm_block_page_add_dynamic_path¶

This directive is used to initialize the blocking page that has NGINX variables in its code and the path to this blocking page is also set using a variable. Otherwise, the directive is not used.

More details on the blocking page and error code configuration →

wallarm_cache_path¶

A directory in which the backup catalog for the proton.db and custom ruleset file copy storage is created when the server starts. This directory must be writable for the client that runs NGINX.

wallarm_custom_ruleset_path¶

A path to the custom ruleset file that contains information on the protected application and the filtering node settings.

wallarm_enable_apifw¶

The directive enables on / disables off API Specification Enforcement, available from release 4.10 onwards. Please note that activating this feature does not substitute for the required subscription and configuration through the Wallarm Console UI.

wallarm_enable_libdetection¶

Enables/disables additional validation of the SQL Injection attacks via the libdetection library. Using libdetection ensures the double‑detection of attacks and reduces the number of false positives.

Analyzing of requests with the libdetection library is enabled by default in all deployment options. To reduce the number of false positives, we recommend analysis to stay enabled.

More details on libdetection →

wallarm_fallback¶

With the value set to on, NGINX has the ability to enter an emergency mode; if proton.db or custom ruleset cannot be downloaded, this setting disables the Wallarm module for the http, server, and location blocks, for which the data fails to download. NGINX keeps functioning.

wallarm_file_check_interval¶

Defines an interval between checking new data in proton.db and custom ruleset file. The unit of measure is specified in the suffix as follows:

• no suffix for minutes

• s for seconds

• ms for milliseconds

wallarm_force¶

Sets the requests' analysis and custom rules generation based on the NGINX mirrored traffic. See Analyzing mirrored traffic with NGINX.

wallarm_general_ruleset_memory_limit¶

Set a limit for the maximum amount of memory that can be used by one instance of proton.db and custom ruleset.

If the memory limit is exceeded while processing some request, the user will get a 500 error.

The following suffixes can be used in this parameter:

• k or K for kilobytes

• m or M for megabytes

• g or G for gigabyte

Value of 0 turns the limit off.

wallarm_global_trainingset_path¶

wallarm_http_v2_stream_max_len¶

This directive sets the maximum allowed length of an HTTP/2 stream in bytes. When half of the specified value is reached, an HTTP/2 GOAWAY frame is sent to the client to facilitate graceful stream termination. If the stream is not closed and the maximum length is reached, the connection is forcefully terminated by NGINX.

If this option is not configured, stream lengths remain unlimited, potentially causing unbounded memory consumption by the NGINX process, particularly in gRPC environments with long-lived connections.

wallarm_instance¶

wallarm_key_path¶

A path to the Wallarm private key used for encryption/decryption of proton.db and custom ruleset files.

wallarm_local_trainingset_path¶

wallarm_memlimit_debug¶

This directive determines whether the Wallarm NGINX module generates the /tmp/proton_last_memlimit.req file containing request details when a memory limit is exceeded. This can be invaluable for debugging issues related to request memory limit processing.

wallarm_mode¶

Traffic processing mode:

• off

• monitoring

• safe_blocking

• block

Usage of wallarm_mode can be restricted by the wallarm_mode_allow_override directive.

Detailed instructions on filtration mode configuration →

wallarm_mode_allow_override¶

Manages the ability to override the wallarm_mode values via filtering rules downloaded from the Wallarm Cloud (custom ruleset):

• off - the custom rules are ignored.
• strict - custom rules can only strengthen the operation mode.
• on - it is possible to both strengthen and soften the operation mode.

For example, with wallarm_mode monitoring and wallarm_mode_allow_override strict set, Wallarm Console can be used to enable blocking of some requests, but the attack analysis cannot be fully disabled.

Detailed instructions on filtration mode configuration →

wallarm_parse_response¶

Whether to analyze the application responses. Response analysis is required for vulnerability detection during passive detection and active threat verification.

Possible values are on (response analysis is enabled) and off (response analysis is disabled).

wallarm_parse_websocket ¶

Wallarm provides full WebSockets support under the API Security subscription plan. By default, the WebSockets' messages are not analyzed for attacks.

To force the feature, activate the API Security subscription plan and use the wallarm_parse_websocket directive.

Possible values:

• on: message analysis is enabled.
• off: message analysis is disabled.

wallarm_parser_disable¶

Allows to disable parsers. The directive values corresponds to the name of the parser to be disabled:

• cookie
• zlib
• htmljs
• json
• multipart
• base64
• percent
• urlenc
• xml
• jwt

Example

wallarm_parser_disable base64;
wallarm_parser_disable xml;
location /ab {
    wallarm_parser_disable json;
    wallarm_parser_disable base64;
    proxy_pass http://example.com;
}
location /zy {
    wallarm_parser_disable json;
    proxy_pass http://example.com;
}

wallarm_parse_html_response¶

Whether to apply the HTML parsers to HTML code received in the application response. Possible values are on (HTML parser is applied) and off (HTML parser is not applied).

This parameter is effective only if wallarm_parse_response on.

wallarm_partner_client_uuid¶

Unique identifier of the tenant for the multi-tenant Wallarm node. The value should be a string in the UUID format, for example:

• 11111111-1111-1111-1111-111111111111

• 123e4567-e89b-12d3-a456-426614174000

Configuration example:

server {
  server_name  tenant1.com;
  wallarm_partner_client_uuid 11111111-1111-1111-1111-111111111111;
  ...
  location /login {
     wallarm_application 21;
     ...
  }
  location /users {
     wallarm_application 22;
     ...
  }

server {
  server_name  tenant1-1.com;
  wallarm_partner_client_uuid 11111111-1111-1111-1111-111111111111;
  wallarm_application 23;
  ...
}

server {
  server_name  tenant2.com;
  wallarm_partner_client_uuid 22222222-2222-2222-2222-222222222222;
  ...
}
...
}

In the configuration above:

• Tenant stands for partner's client. The partner has 2 clients.

• The traffic targeting tenant1.com and tenant1-1.com will be associated with the client 11111111-1111-1111-1111-111111111111.

• The traffic targeting tenant2.com will be associated with the client 22222222-2222-2222-2222-222222222222.

• The first client also has 3 applications, specified via the wallarm_application directive:

  • tenant1.com/login – wallarm_application 21
  • tenant1.com/users – wallarm_application 22
  • tenant1-1.com – wallarm_application 23

  The traffic targeting these 3 paths will be associated with the corresponding application, the remaining will be the generic traffic of the first client.

wallarm_process_time_limit¶

Sets the time limit of a single request processing by the Wallarm node.

If the time exceeds the limit, an error is recorded into the log and the request is marked as the overlimit_res attack. Depending on the wallarm_process_time_limit_block value, the attack can be either blocked, monitored or ignored.

The value is specified in milliseconds without units, e.g.:

wallarm_process_time_limit 1200; # 1200 milliseconds
wallarm_process_time_limit 2000; # 2000 milliseconds

wallarm_process_time_limit_block¶

The ability to manage the blocking of requests, which exceed the time limit set in the wallarm_process_time_limit directive:

• on: the requests are always blocked unless wallarm_mode off

• off: the requests are always ignored

  Protection bypass risk

  The off value should be used carefully as this value disables protection from the overlimit_res attacks.

  It is recommended to use the off value only in the strictly specific locations where it is really necessary, for example where the upload of the large files is performed, and where there is no risk of protection bypass and vulnerability exploit.

  It is strongly not recommended to set wallarm_process_time_limit_block to off globally for http or server blocks.

• attack: depends on the attack blocking mode set in the wallarm_mode directive:

  • off: the requests are not processed.
  • monitoring: the requests are ignored but details on the overlimit_res attacks are uploaded to the Wallarm Cloud and displayed in Wallarm Console.
  • safe_blocking: only requests originated from graylisted IP addresses are blocked and details on all overlimit_res attacks are uploaded to the Wallarm Cloud and displayed in Wallarm Console.
  • block: the requests are blocked.

Regardless of the directive value, requests of the overlimit_res attack type are uploaded to the Wallarm Cloud except when wallarm_mode off;.

wallarm_proton_log_mask_master¶

Settings for the debug logging of the NGINX master process.

wallarm_proton_log_mask_worker¶

Settings of the debug logging for a NGINX worker process.

wallarm_protondb_path¶

A path to the proton.db file that has the global settings for request filtering, which do not depend on the application structure.

wallarm_rate_limit¶

Sets the rate limiting configuration in the following format:

wallarm_rate_limit <KEY_TO_MEASURE_LIMITS_FOR> rate=<RATE> burst=<BURST> delay=<DELAY>;

• KEY_TO_MEASURE_LIMITS_FOR - a key that you want to measure limits for. Can contain text, NGINX variables and their combination.

  For example: "$remote_addr +login" to limit requests originating from the same IP and targeted at the /login endpoint.

• rate=<RATE> (required) - rate limit, can be rate=<number>r/s or rate=<number>r/m.

• burst=<BURST> (optional) - maximum number of excessive requests to be buffered once the specified RPS/RPM is exceeded and to be processed once the rate is back to normal. 0 by default.

• delay=<DELAY> - if the <BURST> value is different from 0, you can control whether to keep the defined RPS/RPM between buffered excessive requests execution. nodelay points to simultaneous processing of all buffered excessive requests, without the rate limit delay. Numeric value implies simultaneous processing of the specified number of excessive requests, others are processed with delay set in RPS/RPM.

Example:

wallarm_rate_limit "$remote_addr +location_name" rate=10r/s burst=9 delay=5;

wallarm_rate_limit_enabled¶

Enables/disables Wallarm rate limiting.

If off, neither the rate limiting rule (recommended) nor the wallarm_rate_limit directive work.

wallarm_rate_limit_log_level¶

The level for logging the requests rejected by the rate limting control. Can be: info, notice, warn, error.

wallarm_rate_limit_status_code¶

Code to return in response to the requests rejected by the Wallarm rate limiting module.

wallarm_rate_limit_shm_size¶

Sets the maximum amount of shared memory that the Wallarm rate limiting module can consume.

With an average key length of 64 bytes (characters), and wallarm_rate_limit_shm_size of 64MB, the module can handle approximately 130,000 unique keys simultaneously. Increasing the memory by two doubles the module's capacity in a linear fashion.

A key is a unique value of a request point that the module uses to measure limits. For instance, if the module is limiting connections based on IP addresses, each unique IP address is considered a single key. With the default directive value, the module can process requests originating from ~130,000 different IPs simultaneously.

wallarm_request_chunk_size¶

Limits the size of the part of the request that is processed during one iteration. You can set up a custom value of the wallarm_request_chunk_size directive in bytes by assigning an integer to it. The directive also supports the following postfixes:

• k or K for kilobytes

• m or M for megabytes

• g or G for gigabytes

wallarm_request_memory_limit¶

Set a limit for the maximum amount of memory that can be used for processing of a single request.

If the limit is exceeded, the request processing will be interrupted and a user will get a 500 error.

The following suffixes can be used in this parameter:

• k or K for kilobytes

• m or M for megabytes

• g or G for gigabytes

Value of 0 turns the limit off.

By default, limits are off.

wallarm_srv_include¶

Specifies the path to the configuration file for API Specification Enforcement. This file is included with all deployment artifacts by default, and usually no changes are needed.

However, if you use an NGINX-based Docker image with a custom nginx.conf, you must specify this directive and place the file at the specified path.

The directive is available from release 4.10.7 onwards.

wallarm_stalled_worker_timeout¶

Sets the time limit for processing a single request for an NGINX worker in seconds.

If the time exceeds the limit, data about NGINX workers is written to the stalled_workers_count and stalled_workers statistic parameters.

wallarm_status¶

Controls the Wallarm statistics service operation.

The directive value has the following format:

wallarm_status [on|off] [format=json|prometheus];

It is highly recommended to configure the statistics service in its own file, avoiding the wallarm_status directive in other NGINX setup files, because the latter may be insecure. The configuration file for wallarm-status is located at:

• /etc/nginx/wallarm-status.conf for all-in-one installer

• /etc/nginx/conf.d/wallarm-status.conf for other installations

Also, it is strongly advised not to alter any of the existing lines of the default wallarm-status configuration as it may corrupt the process of metric data upload to the Wallarm cloud.

wallarm_tarantool_upstream¶

With the wallarm_tarantool_upstream, you can balance the requests between several postanalytics servers.

Example:

upstream wallarm_tarantool {
    server 127.0.0.1:3313 max_fails=0 fail_timeout=0 max_conns=1;
    keepalive 1;
}

# omitted

wallarm_tarantool_upstream wallarm_tarantool;

See also Module ngx_http_upstream_module.

wallarm_timeslice¶

A limit on the time that a filtering node spends on one iteration of processing a request before it switches to the next request. Upon reaching the time limit, the filtering node proceeds to process the next request in the queue. After performing one iteration on each of the requests in the queue, the node performs the second iteration of processing on the first request in the queue.

You can use time intervals suffixes that are described in the NGINX documentation to assign different time unit values to the directive.

wallarm_ts_request_memory_limit¶

wallarm_unpack_response¶

Whether to decompress compressed data returned in the application response. Possible values are on (decompression is enabled) and off (decompression is disabled).

This parameter is effective only if wallarm_parse_response on.

wallarm_upstream_backend¶

A method for sending serialized requests. Requests can be sent either to Tarantool or to the API.

Possible values of the directive:

• tarantool

• api

Depending on the other directives, the default value will be assigned as follows:

• tarantool - if there is no wallarm_api_conf directive in the configuration.

• api - if there is a wallarm_api_conf directive, but there is no wallarm_tarantool_upstream directive in the configuration.

  Note

  If the wallarm_api_conf and wallarm_tarantool_upstream directives are present simultaneously in the configuration, a configuration error of the directive ambiguous wallarm upstream backend form will occur.

wallarm_upstream_connect_attempts¶

Defines the number of immediate reconnects to the Tarantool or Wallarm API.
If a connection to the Tarantool or API is terminated, then the attempt to reconnect will not occur. However, this is not the case when there aren't anymore connections and the serialized request queue is not empty.

wallarm_upstream_reconnect_interval¶

Defines the interval between attempts to reconnect to the Tarantool or Wallarm API after the number of unsuccessful attempts has exceeded the wallarm_upstream_connect_attempts threshold.

wallarm_upstream_connect_timeout¶

Defines a timeout for connecting to the Tarantool or Wallarm API.

wallarm_upstream_queue_limit¶

Defines a limit to the number of serialized requests.
Simultaneously setting the wallarm_upstream_queue_limit parameter and not setting the wallarm_upstream_queue_memory_limit parameter means that there will be no limit on the latter.

wallarm_upstream_queue_memory_limit¶

Defines a limit to the total volume of serialized requests.
Simultaneously setting the wallarm_upstream_queue_memory_limit parameter and not setting the wallarm_upstream_queue_limit parameter means that there will be no limit on the latter.

Was this page helpful?

How can we improve this article? (Optional)

Information is unclear or confusing

Insufficient information

Outdated or incorrect information

0/240

Send