feat: Add support for filtering username/password URL authority. (#751)

Author: kinyoklion

packages/telemetry/browser-telemetry/__tests__/filters/defaultUrlFilter.test.ts

@@ -39,3 +39,27 @@ it.each([
 ])('passes through other URLs unfiltered', (url) => {
   expect(defaultUrlFilter(url)).toBe(url);
 });
+
+it('filters out username and password from URLs', () => {
+  const urls = [
+    // Username only
+    {
+      input: 'https://[email protected]/',
+      expected: 'https://[email protected]/',
+    },
+    // Password only
+    {
+      input: 'https://:[email protected]/',
+      expected: 'https://:[email protected]/',
+    },
+    // Both username and password
+    {
+      input: 'https://user:[email protected]/',
+      expected: 'https://redacted:[email protected]/',
+    },
+  ];
+
+  urls.forEach(({ input, expected }) => {
+    expect(defaultUrlFilter(input)).toBe(expected);
+  });
+});

packages/telemetry/browser-telemetry/src/filters/defaultUrlFilter.ts

@@ -1,13 +1,43 @@
 const pollingRegex = /sdk\/evalx\/[^/]+\/contexts\/(?<context>[^/?]*)\??.*?/;
 const streamingREgex = /\/eval\/[^/]+\/(?<context>[^/?]*)\??.*?/;
 
+/**
+ * Filter which redacts user information (auth) from a URL.
+ *
+ * If a username/password is present, then they are replaced with 'redacted'.
+ * Authority reference: https://developer.mozilla.org/en-US/docs/Web/URI/Authority
+ *
+ * @param url URL to filter.
+ * @returns A filtered URL.
+ */
+function authorityUrlFilter(url: string): string {
+  // This will work in browser environments, but in the future we may want to consider an approach
+  // which doesn't rely on the browser's URL parsing. This is because other environments we may
+  // want to target, such as ReactNative, may not have as robust URL parsing.
+  const urlObj = new URL(url);
+  let hadAuth = false;
+  if (urlObj.username) {
+    urlObj.username = 'redacted';
+    hadAuth = true;
+  }
+  if (urlObj.password) {
+    urlObj.password = 'redacted';
+    hadAuth = true;
+  }
+  if (hadAuth) {
+    return urlObj.toString();
+  }
+  // If there was no auth information, then we don't need to modify the URL.
+  return url;
+}
+
 /**
  * Filter which removes context information for browser JavaScript endpoints.
  *
  * @param url URL to filter.
  * @returns A filtered URL.
  */
-export default function defaultUrlFilter(url: string): string {
+function ldUrlFilter(url: string): string {
   // TODO: Maybe we consider a way to identify LD requests so they can be filtered without
   // regular expressions.
 
@@ -27,3 +57,13 @@ export default function defaultUrlFilter(url: string): string {
   }
   return url;
 }
+
+/**
+ * Filter which redacts user information and removes context information for browser JavaScript endpoints.
+ *
+ * @param url URL to filter.
+ * @returns A filtered URL.
+ */
+export default function defaultUrlFilter(url: string): string {
+  return ldUrlFilter(authorityUrlFilter(url));
+}