url obfuscation

newrelic/api/web_transaction.py

@@ -16,6 +16,7 @@
 import time
 import logging
 import warnings
+import re
 
 try:
     import urlparse
@@ -139,6 +140,23 @@ class WebTransaction(Transaction):
     unicode_error_reported = False
     QUEUE_TIME_HEADERS = ('x-request-start', 'x-queue-start')
 
+    @property
+    def _request_uri(self):
+        return self.request_uri
+
+
+    @_request_uri.setter
+    def _request_uri(self, raw_uri):
+        if self._settings and raw_uri:
+            if self._settings.url_regex and self._settings.url_regex_rep:
+                obfuscated_uri = re.sub(str(self._settings.url_regex),str(self._settings.url_regex_rep), str(raw_uri))
+                self.request_uri = obfuscated_uri
+            elif self._settings.url_regex:
+                obfuscated_uri = re.sub(str(self._settings.url_regex),"*****", str(raw_uri))
+                self.request_uri = obfuscated_uri
+            else:
+                self.request_uri = raw_uri
+
     def __init__(self, application, name, group=None,
             scheme=None, host=None, port=None, request_method=None,
             request_path=None, query_string=None, headers=None,

newrelic/config.py

@@ -314,6 +314,8 @@ def _process_setting(section, option, getter, mapper):
 def _process_configuration(section):
     _process_setting(section, "feature_flag", "get", _map_feature_flag)
     _process_setting(section, "app_name", "get", None)
+    _process_setting(section, "url_regex", "get", None)
+    _process_setting(section, "url_regex_rep", "get", None)
     _process_setting(section, "labels", "get", _map_labels)
     _process_setting(section, "license_key", "get", _map_default_host_value)
     _process_setting(section, "api_key", "get", None)

newrelic/core/config.py

@@ -570,6 +570,9 @@ def default_host(license_key):
 _settings.ca_bundle_path = os.environ.get("NEW_RELIC_CA_BUNDLE_PATH", None)
 
 _settings.app_name = os.environ.get("NEW_RELIC_APP_NAME", "Python Application")
+_settings.url_regex = os.environ.get('NEW_RELIC_URL_REGEX', None )
+_settings.url_regex_rep = os.environ.get('NEW_RELIC_URL_REGEX_REP', None)
+
 _settings.linked_applications = []
 
 _settings.process_host.display_name = os.environ.get("NEW_RELIC_PROCESS_HOST_DISPLAY_NAME", None)

newrelic/newrelic.ini

@@ -42,6 +42,14 @@ license_key = *** REPLACE ME ***
 # https://docs.newrelic.com/docs/apm/agents/manage-apm-agents/app-naming/use-multiple-names-app/
 app_name = Python Application
 
+# If sensitive data needs obfuscating in the URL, set this to regex pattern to search in URL
+# Disabled by default
+url_regex = 
+
+# String to replace with if regex pattern returns match
+# Default *****
+url_regex_rep = 
+
 # When "true", the agent collects performance data about your
 # application and reports this data to the New Relic UI at
 # newrelic.com. This global switch is normally overridden for