Overview

If you experience unexpected behavior with Datadog Application Security Management (ASM), there are common issues you can investigate, as mentioned below. If you continue to have trouble, reach out to Datadog support for further assistance.

ASM rate limits

ASM traces are rate-limited to 100 traces per second. Traces sent after the limit are not reported. Contact Datadog support if you need to change the limit.

No security traces detected by ASM

There are a series of steps that must run successfully for threat information to appear in the ASM Trace and Signals Explorer. It is important to check each step when investigating this issue. Additional troubleshooting steps for specific languages are in the language tab at the end.

Confirm ASM is enabled

You can use the metric datadog.apm.appsec_host to check if ASM is running.

  1. Go to Metrics > Summary in Datadog.
  2. Search for the metric datadog.apm.appsec_host. If the metric doesn’t exist, then there are no services running ASM. If the metric exists, the services are reported with the metric tags host and service.
  3. Select the metric, and in the Tags section, search for service to see which services are running ASM.

If you are not seeing datadog.apm.appsec_host, check the in-app instructions to confirm that all steps for the initial setup are complete.

ASM data is sent with APM traces. See APM troubleshooting to confirm APM setup and check for connection errors.

Send a test attack to your application

To test your ASM setup, trigger the Security Scanner Detected rule by running a file that contains the following curl script:

for ((i=1;i<=250;i++));
do
# Target existing service's routes
curl https://your-application-url/existing-route -A dd-test-scanner-log;
# Target non existing service's routes
curl https://your-application-url/non-existing-route -A dd-test-scanner-log;
done

Note: The dd-test-scanner-log value is supported in the most recent releases.

for ((i=1;i<=250;i++));
do
# Target existing service's routes
curl https://your-application-url/existing-route -A dd-test-scanner-log;
# Target non existing service's routes
curl https://your-application-url/non-existing-route -A dd-test-scanner-log;
done

Note: The dd-test-scanner-log value is supported in the most recent releases.

for ((i=1;i<=250;i++));
do
# Target existing service's routes
curl https://your-application-url/existing-route -A Arachni/v1.0;
# Target non existing service's routes
curl https://your-application-url/non-existing-route -A Arachni/v1.0;
done
for ((i=1;i<=250;i++));
do
# Target existing service's routes
curl https://your-application-url/existing-route -A Arachni/v1.0;
# Target non existing service's routes
curl https://your-application-url/non-existing-route -A Arachni/v1.0;
done
for ((i=1;i<=250;i++));
do
# Target existing service's routes
curl https://your-application-url/existing-route -A dd-test-scanner-log;
# Target non existing service's routes
curl https://your-application-url/non-existing-route -A dd-test-scanner-log;
done

Note: The dd-test-scanner-log value is supported in the most recent releases.

for ((i=1;i<=250;i++));
do
# Target existing service's routes
curl https://your-application-url/existing-route -A dd-test-scanner-log;
# Target non existing service's routes
curl https://your-application-url/non-existing-route -A dd-test-scanner-log;
done

Note: The dd-test-scanner-log value is supported in the most recent releases.

for ((i=1;i<=250;i++));
do
# Target existing service's routes
curl https://your-application-url/existing-route -A dd-test-scanner-log;
# Target non existing service's routes
curl https://your-application-url/non-existing-route -A dd-test-scanner-log;
done

A few minutes after you enable your application and exercise it, and if it’s successful, threat information appears in the Trace and Signals Explorer.

Security Signal details page showing tags, metrics, suggested next steps, and attacker IP addresses associated with a threat.

Check if required tracer integrations are deactivated

ASM relies on certain tracer integrations. If they are deactivated, ASM won’t work. To see if there are deactivated integrations, look for disabled_integrations in your startup logs.

The required integrations vary by language.

For Java, if you are using any of the following technologies, the respective integration is required:

  • grizzly
  • grizzly-filterchain
  • jersey
  • jetty
  • ratpack
  • ratpack-request-body (also requires ratpack)
  • resteasy
  • servlet
  • servlet-2
  • servlet-3
  • servlet-request-body (also requires servlet)
  • spring-web
  • tomcat

For .NET, the ASP.NET integration is required.

Note: If ASP.NET Core is disabled, ASM should still work with this framework.

There are no required integrations for PHP.

The following Go frameworks should be instrumented using the out-of-the-box APM integrations:

If your framework is not supported, create a new issue in the Go repository.

For Node.js, the HTTP integration is required.

For Ruby, the Rack integration is required. Ruby tracer version 1.0.0 or higher is also required. See information on migrating from 0.x to 1.x.

Note: Rack can be manually added or automatically added with the Rails or Sinatra integration. If manually added, the tracer middleware must appear before the security middleware in the Rack stack.

For Python, the WSGI integration is required along with the integration for the framework you’re using, such as the Django or Flask integration.

Check Datadog Agent configuration

To troubleshoot this step of the process, do the following:

  • Check the details of the running Agent at this address http://<agent-machine-name>:<agent-port>/info, usually http://localhost:8126/info.
  • Ensure there are no Agent transmission errors related to spans in your tracer logs.
  • If the Agent is installed on a separate machine, check that DD_AGENT_HOST and, optionally, DD_TRACE_AGENT_PORT are set, or that DD_TRACE_AGENT_URL is set for the application tracing library.

Check if spans are successfully transmitted to Datadog

ASM data is sent over spans. To confirm that spans are successfully transmitted to Datadog, check that your tracer logs contain logs that look similar to this:

2021-11-29 21:19:58 CET | TRACE | INFO | (pkg/trace/info/stats.go:111 in LogStats) | [lang:.NET lang_version:5.0.10 interpreter:.NET tracer_version:1.30.1.0 endpoint_version:v0.4] -> traces received: 2, traces filtered: 0, traces amount: 1230 bytes, events extracted: 0, events sampled: 0

If spans are not being transmitted, then the tracer logs will contain logs similar to this:

2021-11-29 21:18:48 CET | TRACE | INFO | (pkg/trace/info/stats.go:104 in LogStats) | No data received

Troubleshooting by language

Below are additional troubleshooting steps for specific languages.

The Java library uses SLF4J for logging. Add the following runtime flags so that the tracer logs to a file:

 -Ddatadog.slf4j.simpleLogger.defaultLogLevel=info
 -Ddatadog.slf4j.simpleLogger.logFile=dd.log

After the service starts, the tracer logs to the specified file. Datadog recommends using INFO for the log level because DEBUG logs are verbose.

The .NET library logs to a file and cannot log to stdout/stderr. The default log level is INFO. To enable DEBUG logs, set DD_TRACE_DEBUG=true.

The log files are available in the following directories:

PlatformLog directory
DockerThe container’s directory /var/log/datadog/dotnet/. A recommended option is to mount the log folder on the host machine using volumes.
Linux/var/log/datadog/dotnet/
WindowsC:\ProgramData\Datadog .NET Tracer\logs

For PHP, to start troubleshooting issues with the Datadog ASM extension, enable debug logs in the ASM extension’s .ini file.

The extension’s ini file is usually found in /etc/php/<version>/xxx/conf.d/98-ddtrace.ini, but the location may differ depending on your installation. Look at the beginning of the phpinfo() output to identify the directory that is scanned for .ini files, if any. In the .ini file, set the following configuration options with the following:

datadog.appsec.log_level='debug'
datadog.appsec.helper_extra_args='--log_level=debug'
datadog.appsec.helper_log_file='/tmp/helper.log'

The extension outputs logs to the default php_error log file. If there are no logs in the file, add the following to the .ini file:

datadog.appsec.log_file='tmp/extension.log'

Installation fails to find PHP

If the installation script is unable to find the correct PHP version, you can set the --php-bin to the PHP binary location, for example:

$ php datadog-setup.php --php-bin /usr/bin/php7.4 --enable-appsec

Connection to helper failed

If the ASM extension is unable to communicate with the helper process, the following warning occurs:

PHP Warning:  Unknown: [ddappsec] Connection to helper failed and we are not going to attempt to launch it: dd_error

The warning could be followed by one of these error messages:

PHP Warning:  Unknown: [ddappsec] Could not open lock file /tmp/ddappsec.lock: Permission denied in Unknown on line 0
PHP Warning:  Unknown: [ddappsec] Call to bind() failed: Permission denied
PHP Warning:  Unknown: [ddappsec] Failed to unlink /tmp/ddappsec.sock: Operation not permitted

This indicates that the lock file or socket file used by the extension has invalid permissions, or the user executing the PHP process does not have write access to the tmp directory.

If the lock file or socket file has invalid permissions, you can either delete them and restart Apache/FPM or adjust the user:group to match the one used by Apache/FPM, for example, www-data.

If the user doesn’t have write access to the tmp directory, you can change the location of the lock file and socket file by modifying the following settings in the extension’s .ini file:

datadog.appsec.helper_runtime_path = /<directory with compatible permissions>/

Confirm ASM is enabled in the running application

Tracer startup logs show the tracer configuration and whether ASM is enabled or not. If appsec is true, then ASM is enabled and running.

For example, the following startup log shows that ASM is disabled:

2022/02/17 14:49:00 Datadog Tracer v1.36.0 INFO: DATADOG TRACER CONFIGURATION {"date":"2022-02-17T14:49:00+01:00","os_name":"Linux (Unknown Distribution)","os_version":"5.13.0","version":"v1.36.0","lang":"Go","lang_version":"go1.17.1","env":"prod","service":"grpcserver","agent_url":"http://localhost:8126/v0.4/traces","debug":false,"analytics_enabled":false,"sample_rate":"NaN","sampling_rules":null,"sampling_rules_error":"","service_mappings":null,"tags":{"runtime-id":"69d99219-b68f-4718-9419-fa173a79351e"},"runtime_metrics_enabled":false,"health_metrics_enabled":false,"profiler_code_hotspots_enabled":false,"profiler_endpoints_enabled":false,"dd_version":"","architecture":"amd64","global_service":"","lambda_mode":"false","appsec":false,"agent_features":{"DropP0s":false,"Stats":false,"StatsdPort":0}}

Enable debug logs

Enable debug logs with the environment variable DD_TRACE_DEBUG=1. The ASM library will log to the standard error output.

Note: ASM only outputs logs when it is enabled. Use the environment variable DD_APPSEC_ENABLED=1 to enable ASM.

Use this migration guide to assess any breaking changes if you upgraded your Node.js library from 1.x to 2.x.

If you don’t see ASM threat information in the Trace and Signals Explorer for your Node.js application, follow these steps to troubleshoot the issue:

  1. Confirm the latest version of ASM is running by checking that appsec_enabled is true in the startup logs

    a. If you don’t see startup logs after a request has been sent, add the environment variable DD_TRACE_STARTUP_LOGS=true to enable startup logs. Check the startup logs for appsec_enabled is true.

    b. If appsec_enabled is false, then ASM was not enabled correctly. See installation instructions.

    c. If appsec_enabled is not in the startup logs, the latest ASM version needs to be installed. See installation instructions.

  2. Is the tracer working? Can you see relevant traces on the APM dashboard?

    ASM relies on the tracer so if you don’t see traces, then the tracer might not be working. See APM Troubleshooting.

  3. In your application directory, run the command npm explore @datadog/native-appsec -- npm run install and restart your app.

    a. If @datadog/native-appsec is not found then the installation is incorrect. See installation instructions.

    b. If @datadog/native-appsec is found when starting your application, add the command to your runtime start script.

    c. If the tracer still does not work, you might be running an unsupported runtime.

  4. To enable logs, add the following environment variables:

    DD_TRACE_DEBUG=1
    DD_TRACE_LOG_LEVEL=info
    

If you don’t see ASM threat information in the Trace and Signals Explorer for your Python application, check that ASM is running and that your tracer is working.

  1. Set your application’s log level to DEBUG to confirm that ASM is running:

    import logging
    logging.basicConfig(level=logging.DEBUG)
    

    Then, run any HTTP call to the application. You should see the following log:

    DEBUG:ddtrace.appsec.processor:[DDAS-001-00] Executing AppSec In-App WAF with parameters:
    

    If this log is not present, ASM is not running.

  2. Is the tracer working? Can you see relevant traces on the APM dashboard?

    ASM relies on the tracer. If you don’t see traces, then the tracer might not be working. See APM Troubleshooting.

For Ruby, if you don’t see ASM threat information in the Trace and Signals Explorer after a few minutes, enable tracer diagnostics for debug logs. For example:

Datadog.configure do |c|
  c.diagnostics.debug = true  # increase general log level to debug
  c.appsec.waf_debug = true   # also enable WAF-specific log verbosity to highest level
end

Debug logs are verbose but useful. If you open up a ticket with Datadog support, forward the logs with your request.

Is ASM correctly enabled?

ASM has been correctly enabled if you see logs such as:

D, [2021-12-14T11:03:32.167125 #73127] DEBUG -- ddtrace: [ddtrace] (libddwaf/lib/datadog/appsec/waf.rb:296:in `block in logger=') {:level=>:ddwaf_log_info, :func=> "ddwaf_set_log_cb", :file=>"PowerWAFInterface.cpp", :message=>"Sending log messages to binding, min level trace"}
D, [2021-12-14T11:03:32.200491 #73127] DEBUG -- ddtrace: [ddtrace] (libddwaf/lib/datadog/appsec/waf.rb:296:in `block in logger=') {:level=>:ddwaf_log_debug, :func= >"parse", :file=>"parser_v2.cpp", :message=>"Loaded 124 rules out of 124 available in the ruleset"}

If you do not see those logs, check the following:

  • If the correct ASM environment variables are set for your application process.
  • The latest gem version is installed.
  • The tracer is configured correctly and sending APM traces to your APM dashboard.

Is ASM called for each HTTP request?

To confirm that ASM is called for each HTTP request, trigger a test attack and look for these logs:

D, [2022-01-19T21:25:50.579745 #341792] DEBUG -- ddtrace: [ddtrace] (/home/lloeki/src/github.com/DataDog/dd-trace-rb/lib/datadog/appsec/reactive/operation.rb:14:in `initialize') operation: rack.request initialize
D, [2022-01-19T21:25:50.580300 #341792] DEBUG -- ddtrace: [ddtrace] (/home/lloeki/src/github.com/DataDog/dd-trace-rb/lib/datadog/appsec/contrib/rack/gateway/watcher.rb:25:in `block (2 levels) in watch') root span: 964736568335365930
D, [2022-01-19T21:25:50.580371 #341792] DEBUG -- ddtrace: [ddtrace] (/home/lloeki/src/github.com/DataDog/dd-trace-rb/lib/datadog/appsec/contrib/rack/gateway/watcher.rb:26:in `block (2 levels) in watch') active span: 964736568335365930
D, [2022-01-19T21:25:50.581061 #341792] DEBUG -- ddtrace: [ddtrace] (/home/lloeki/src/github.com/DataDog/dd-trace-rb/lib/datadog/appsec/contrib/rack/reactive/request.rb:34:in `block in subscribe') reacted to ["request.headers", "request.uri.raw", "request.query", "request.cookies", "request.body.raw"]: [{"version"=>"HTTP/1.1", "host"=>"127.0.0.1:9292", "accept"=>"*/*", "user-agent"=>"Nessus SOAP"}, "http://127.0.0.1:9292/", [], {}, ""]

If you don’t see those logs, try the following:

  • Check that another upstream security system is not filtering requests based on the test header value, which would prevent the request from reaching the application.
  • Send another test attack using another user agent value in the curl command to see if the threat information is successfully sent.
  • Look in the application logs for the exact request you ran to confirm the request reached the application, and was not responded to by another upstream system.

If the Rack integration was configured manually, sometimes a known issue prevents ASM from working. For example:

Datadog.configure do |c|
  c.tracing.instrument :rails
  ...
  c.tracing.instrument :rack, web_service_name: "something", request_queuing: true

If c.tracing.instrument :rack is present, remove it to see if the check passes.

Is ASM detecting HTTP request security threats?

To confirm that ASM is detecting security threats, trigger a test attack, and look for these logs:

D, [2021-12-14T22:39:53.268820 #106051] DEBUG -- ddtrace: [ddtrace] (ddtrace/lib/datadog/appsec/contrib/rack/reactive/request.rb:63:in `block in subscribe') WAF: #<struct Datadog::AppSec::WAF::Result action=:monitor, data=[{"rule"=>{"id"=>"ua0-600-10x", "name"=>"Nessus", "tags"=>{"type"=>"security_scanner", "category"=>"attack_attempt"}}, "rule_matches"=>[{"operator"=>"match_regex", "operator_value"=>"(?i)^Nessus(/|([ :]+SOAP))", "parameters"=>[{"address"=>"server.request.headers.no_cookies", "key_path"=>["user-agent"], "value"=>"Nessus SOAP", "highlight"=>["Nessus SOAP"]}]}]}], perf_data=nil, perf_total_runtime=20519>

If you don’t see those logs, check that another upstream security system is not filtering out the requests or altering them based on the test header value.

Is the tracer sending traces with security data?

ASM data is sent with APM traces. To confirm that ASM correctly detects and inserts security data into traces, trigger a test attack, and look for these tracer logs:

Tags: [
   runtime-id => 0c3dfc67-9cf3-457c-a980-0229b203d048,
   _dd.runtime_family => ruby,
   appsec.event => true,
   _dd.appsec.json => {"triggers":[{"rule":{"id":"ua0-600-10x","name":"Nessus","tags":{"type":"security_scanner","category":"attack_attempt"}},"rule_matches":[{"operator":"match_regex","operator_value":"(?i)^Nessus(/|([ :]+SOAP))","parameters":[{"address":"server.request.headers.no_cookies","key_path":["user-agent"],"value":"Nessus SOAP","highlight":["Nessus SOAP"]}]}]}]},
   http.request.headers.host => 127.0.0.1:9292,
   http.request.headers.accept => */*,
   http.request.headers.user-agent => Nessus SOAP,
   http.response.headers.content-type => text/plain,
   http.host => 127.0.0.1,
   http.useragent => Nessus SOAP,
   network.client.ip => 127.0.0.1,
   _dd.origin => appsec,
   http.method => GET,
   http.url => /,
   http.base_url => http://127.0.0.1:9292,
   http.status_code => 200,
   http.response.headers.content_type => text/plain]
Metrics: [
   _dd.agent_psr => 1.0,
   system.pid => 155644.0,
   _dd.appsec.enabled => 1.0,
   _dd.measured => 1.0,
   _sampling_priority_v1 => 2.0]]

Wait a minute for the agent to forward the traces, then check that the traces show up in the APM dashboard. The security information in the traces may take additional time to be processed by Datadog before showing up as security traces in the ASM Trace and Signals Explorer.

No vulnerabilities detected by Software Composition Analysis

There are a series of steps that must run successfully for vulnerability information to appear either in the Service Catalog Security View or in the Vulnerability Explorer. It is important to check each step when investigating this issue.

Confirm ASM is enabled

You can use the metric datadog.apm.appsec_host to check if ASM is running.

  1. Go to Metrics > Summary in Datadog.
  2. Search for the metric datadog.apm.appsec_host. If the metric doesn’t exist, then there are no services running ASM. If the metric exists, the services are reported with the metric tags host and service.
  3. Select the metric, and in the Tags section, search for service to see which services are running ASM.

If you are not seeing datadog.apm.appsec_host, check the in-app instructions to confirm that all steps for the initial setup are complete.

ASM data is sent with APM traces. See APM troubleshooting to confirm APM setup and check for connection errors.

Confirm tracer versions are updated

See the Application Security set up documentation to validate you you are using the right version of the tracer. These minimum versions are required to start sending telemetry data that includes library information.

Ensure the communication of telemetry data

Ensure the DD_INSTRUMENTATION_TELEMETRY_ENABLED environment variable (DD_TRACE_TELEMETRY_ENABLED for NodeJS) is set to true, or the corresponding system property for your language is enabled. For example in Java: -Ddd.instrumentation.telemetry.enabled=true

Disabling threat management and protection

To disable threat management, remove the DD_APPSEC_ENABLED=true environment variable from your application configuration, and restart your service.

If no DD_APPSEC_ENABLED=true environment variable is set for your service, do one of the following:

  • If it’s a PHP service: explicitly set the environment variable to DD_APPSEC_ENABLED=false, and restart your service.
  • If threat management was activated using Remote Configuration, do the following:
    1. Go to Services (ASM > Catalog > Services).
    2. Select Threat Management in Monitoring Mode.
    3. In the Threat Management facet, enable Monitoring Only, No data, and Ready to block.
    4. Click on a service.
    5. In the service details, in Threat Detection, click Deactivate.
If threat management was activated using Remote Configuration, you can use a Deactivate button. If threat management was activated using local configuration, the Deactivate button is not an option.
  • To disable threat management on your services in bulk, do the following:
    1. Go to Services.
    2. In the Threat Management facet, enable Monitoring Only, No data, and Ready to block.
    3. Select the check boxes for the services where you want to disable threat detection.
    4. In Bulk Actions, select Deactivate Threat detection on (number of) services.

Disabling Software Composition Analysis

To disable Software Composition Analysis:

  • Go to Services, select Software Composition Analysis (SCA), click on your service and then click Deactivate.
  • To disable Software Composition Analysis on your services in bulk, click the check box in the list header and then under Bulk Actions select Deactivate Software Composition Analysis (SCA) on (number of) services.
  • To disable Software Composition Analysis using the DD_APPSEC_SCA_ENABLED environment variable, remove the DD_APPSEC_SCA_ENABLED=true environment variable from your application configuration, and restart your service. This does not apply to PHP apps.

Disabling Code Security

To disable Code Security vulnerability detection, remove the DD_IAST_ENABLED=true environment variable from your application configuration, and restart your service. This does not apply to PHP apps.

If no DD_IAST_ENABLED=true environment variable is set for your service, do the following:

  1. Go to Services.
  2. Select Code Security.
  3. Click on a service.
  4. In the service details, in Vulnerability Detection, click Deactivate.

Some or all IAST/Custom Code vulnerabilities are not detected

Confirm IAST is enabled

Ensure the DD_IAST_ENABLED environment variable is set to true or the corresponding system property for your language is enabled.

For Python+Flask, call the entrypoint patch function

If you’re running a Flask application ensure that you are calling the ddtrace_iast_flask_patch() function at the top level of the module and before calling app.run(). See the Flask integration documentation for more information.

Need more help?

If you continue to have issues with ASM, contact Datadog support with the following information:

Further Reading