Signals sampling (#1166)

Author: danieljackins

.changeset/nasty-cherries-burn.md

@@ -0,0 +1,6 @@
+---
+'@segment/analytics-next': minor
+'@segment/analytics-signals': minor
+---
+
+Add sampling logic and block non debug traffic

packages/browser/src/browser/index.ts

@@ -122,6 +122,13 @@ export interface CDNSettings {
         version: number
       }
     | {}
+
+  /**
+   * Settings for auto instrumentation
+   */
+  autoInstrumentationSettings?: {
+    sampleRate: number
+  }
 }
 
 export interface AnalyticsBrowserSettings {

packages/signals/signals-integration-tests/src/helpers/base-page-object.ts

@@ -74,6 +74,7 @@ export class BasePage {
       ({ signalSettings }) => {
         window.signalsPlugin = new window.SignalsPlugin({
           disableSignalsRedaction: true,
+          enableSignalsIngestion: true,
           ...signalSettings,
         })
         window.analytics.load({

packages/signals/signals-integration-tests/src/tests/signals-vanilla/change-input.test.ts

@@ -14,6 +14,7 @@ test('Collecting signals whenever a user enters text input', async ({
    */
   await indexPage.loadAndWait(page, basicEdgeFn, {
     disableSignalsRedaction: true,
+    enableSignalsIngestion: true,
   })
 
   await Promise.all([

packages/signals/signals-integration-tests/src/tests/signals-vanilla/signals-ingestion.test.ts

@@ -0,0 +1,30 @@
+import { test, expect } from '@playwright/test'
+import { IndexPage } from './index-page'
+
+const indexPage = new IndexPage()
+
+const basicEdgeFn = `const processSignal = (signal) => {}`
+
+test('ingestion not enabled -> will not send the signal', async ({ page }) => {
+  await indexPage.loadAndWait(page, basicEdgeFn, {
+    enableSignalsIngestion: false,
+  })
+
+  await indexPage.fillNameInput('John Doe')
+  await indexPage.waitForSignalsApiFlush().catch(() => {
+    expect(true).toBe(true)
+  })
+})
+
+test('ingestion enabled -> will send the signal', async ({ page }) => {
+  await indexPage.loadAndWait(page, basicEdgeFn, {
+    enableSignalsIngestion: true,
+  })
+
+  await Promise.all([
+    indexPage.fillNameInput('John Doe'),
+    indexPage.waitForSignalsApiFlush(),
+  ])
+
+  expect(true).toBe(true)
+})

packages/signals/signals-integration-tests/src/tests/signals-vanilla/signals-redaction.test.ts

@@ -11,6 +11,7 @@ test('redaction enabled -> will XXX the value of text input', async ({
 }) => {
   await indexPage.loadAndWait(page, basicEdgeFn, {
     disableSignalsRedaction: false,
+    enableSignalsIngestion: true,
   })
 
   await Promise.all([
@@ -40,6 +41,7 @@ test('redation disabled -> will not touch the value of text input', async ({
 }) => {
   await indexPage.loadAndWait(page, basicEdgeFn, {
     disableSignalsRedaction: true,
+    enableSignalsIngestion: true,
   })
 
   await Promise.all([

packages/signals/signals/src/core/client/__tests__/client.test.ts

@@ -11,7 +11,9 @@ describe(SignalsIngestClient, () => {
   let client: SignalsIngestClient
 
   beforeEach(async () => {
-    client = new SignalsIngestClient()
+    client = new SignalsIngestClient({
+      shouldIngestSignals: () => true,
+    })
     await client.init({ writeKey: 'test' })
   })
 

packages/signals/signals/src/core/client/index.ts

@@ -7,22 +7,25 @@ export class SignalsIngestSettings {
   flushAt: number
   flushInterval: number
   apiHost: string
-  shouldDisableSignalRedaction: () => boolean
+  shouldDisableSignalsRedaction: () => boolean
+  shouldIngestSignals: () => boolean
   writeKey?: string
   constructor(settings: SignalsIngestSettingsConfig) {
     this.flushAt = settings.flushAt ?? 5
     this.apiHost = settings.apiHost ?? 'signals.segment.io/v1'
     this.flushInterval = settings.flushInterval ?? 2000
-    this.shouldDisableSignalRedaction =
-      settings.shouldDisableSignalRedaction ?? (() => false)
+    this.shouldDisableSignalsRedaction =
+      settings.shouldDisableSignalsRedaction ?? (() => false)
+    this.shouldIngestSignals = settings.shouldIngestSignals ?? (() => false)
   }
 }
 
 export interface SignalsIngestSettingsConfig {
   apiHost?: string
   flushAt?: number
   flushInterval?: number
-  shouldDisableSignalRedaction?: () => boolean
+  shouldDisableSignalsRedaction?: () => boolean
+  shouldIngestSignals?: () => boolean
 }
 /**
  * This currently just uses the Segment analytics-next library to send signals.
@@ -73,7 +76,10 @@ export class SignalsIngestClient {
     if (!this.analytics) {
       throw new Error('Please initialize before calling this method.')
     }
-    const disableRedaction = this.settings.shouldDisableSignalRedaction()
+    if (!this.settings.shouldIngestSignals()) {
+      return
+    }
+    const disableRedaction = this.settings.shouldDisableSignalsRedaction()
     const cleanSignal = disableRedaction ? signal : redactSignalData(signal)
 
     if (disableRedaction) {

packages/signals/signals/src/core/signals/settings.ts

@@ -14,6 +14,7 @@ export type SignalsSettingsConfig = Pick<
   | 'flushAt'
   | 'flushInterval'
   | 'disableSignalsRedaction'
+  | 'enableSignalsIngestion'
   | 'networkSignalsAllowList'
   | 'networkSignalsDisallowList'
   | 'networkSignalsAllowSameDomain'
@@ -33,7 +34,8 @@ export class SignalGlobalSettings {
   ingestClient: SignalsIngestSettingsConfig
   network: NetworkSettingsConfig
 
-  private redaction = new SignalRedactionSettings()
+  private sampleSuccess = false
+  private signalsDebug = new SignalsDebugSettings()
 
   constructor(settings: SignalsSettingsConfig) {
     if (settings.maxBufferSize && settings.signalStorage) {
@@ -42,8 +44,9 @@ export class SignalGlobalSettings {
       )
     }
 
-    this.redaction = new SignalRedactionSettings(
-      settings.disableSignalsRedaction
+    this.signalsDebug = new SignalsDebugSettings(
+      settings.disableSignalsRedaction,
+      settings.enableSignalsIngestion
     )
 
     this.signalBuffer = {
@@ -54,7 +57,17 @@ export class SignalGlobalSettings {
       apiHost: settings.apiHost,
       flushAt: settings.flushAt,
       flushInterval: settings.flushInterval,
-      shouldDisableSignalRedaction: this.redaction.getDisableSignalRedaction,
+      shouldDisableSignalsRedaction:
+        this.signalsDebug.getDisableSignalsRedaction,
+      shouldIngestSignals: () => {
+        if (this.signalsDebug.getEnableSignalsIngestion()) {
+          return true
+        }
+        if (!this.sampleSuccess) {
+          return false
+        }
+        return false
+      },
     }
     this.sandbox = {
       functionHost: settings.functionHost,
@@ -70,6 +83,7 @@ export class SignalGlobalSettings {
   public update({
     edgeFnDownloadURL,
     disallowListURLs,
+    sampleRate,
   }: {
     /**
      * The URL to download the edge function from
@@ -79,58 +93,80 @@ export class SignalGlobalSettings {
      * Add new URLs to the disallow list
      */
     disallowListURLs: (string | undefined)[]
+    /**
+     * Sample rate to determine sending signals
+     */
+    sampleRate?: number
   }): void {
     edgeFnDownloadURL && (this.sandbox.edgeFnDownloadURL = edgeFnDownloadURL)
     this.network.networkSignalsFilterList.disallowed.addURLLike(
       ...disallowListURLs.filter(<T>(val: T): val is NonNullable<T> =>
         Boolean(val)
       )
     )
+    if (sampleRate && Math.random() <= sampleRate) {
+      this.sampleSuccess = true
+    }
   }
 }
 
-class SignalRedactionSettings {
+class SignalsDebugSettings {
   private static redactionKey = 'segment_signals_debug_redaction_disabled'
-  constructor(initialValue?: boolean) {
-    if (typeof initialValue === 'boolean') {
-      this.setDisableSignalRedaction(initialValue)
+  private static ingestionKey = 'segment_signals_debug_ingestion_enabled'
+  constructor(disableRedaction?: boolean, enableIngestion?: boolean) {
+    if (typeof disableRedaction === 'boolean') {
+      this.setDebugKey(SignalsDebugSettings.redactionKey, disableRedaction)
+    }
+    if (typeof enableIngestion === 'boolean') {
+      this.setDebugKey(SignalsDebugSettings.ingestionKey, enableIngestion)
     }
 
-    // setting ?segment_signals_debug=true will disable redaction, and set a key in local storage
+    // setting ?segment_signals_debug=true will disable redaction, enable ingestion, and set keys in local storage
     // this setting will persist across page loads (even if there is no query string)
     // in order to clear the setting, user must set ?segment_signals_debug=false
     const debugModeInQs = parseDebugModeQueryString()
     logger.debug('debugMode is set to true via query string')
     if (typeof debugModeInQs === 'boolean') {
-      this.setDisableSignalRedaction(debugModeInQs)
+      this.setDebugKey(SignalsDebugSettings.redactionKey, debugModeInQs)
+      this.setDebugKey(SignalsDebugSettings.ingestionKey, debugModeInQs)
     }
   }
 
-  setDisableSignalRedaction(shouldDisable: boolean) {
+  setDebugKey(key: string, enable: boolean) {
     try {
-      if (shouldDisable) {
-        window.sessionStorage.setItem(
-          SignalRedactionSettings.redactionKey,
-          'true'
-        )
+      if (enable) {
+        window.sessionStorage.setItem(key, 'true')
       } else {
-        logger.debug('Removing redaction key from storage')
-        window.sessionStorage.removeItem(SignalRedactionSettings.redactionKey)
+        logger.debug(`Removing debug key ${key} from storage`)
+        window.sessionStorage.removeItem(key)
+      }
+    } catch (e) {
+      logger.debug('Storage error', e)
+    }
+  }
+
+  getDisableSignalsRedaction() {
+    try {
+      const isEnabled = Boolean(
+        window.sessionStorage.getItem(SignalsDebugSettings.redactionKey)
+      )
+      if (isEnabled) {
+        logger.debug(`${SignalsDebugSettings.redactionKey}=true (app. storage)`)
+        return true
       }
     } catch (e) {
       logger.debug('Storage error', e)
     }
+    return false
   }
 
-  getDisableSignalRedaction() {
+  getEnableSignalsIngestion() {
     try {
-      const isDisabled = Boolean(
-        window.sessionStorage.getItem(SignalRedactionSettings.redactionKey)
+      const isEnabled = Boolean(
+        window.sessionStorage.getItem(SignalsDebugSettings.ingestionKey)
       )
-      if (isDisabled) {
-        logger.debug(
-          `${SignalRedactionSettings.redactionKey}=true (app. storage)`
-        )
+      if (isEnabled) {
+        logger.debug(`${SignalsDebugSettings.ingestionKey}=true (app. storage)`)
         return true
       }
     } catch (e) {

packages/signals/signals/src/core/signals/signals.ts

@@ -91,6 +91,9 @@ export class Signals implements ISignals {
         analyticsService.instance.settings.apiHost,
         analyticsService.instance.settings.cdnURL,
       ],
+      sampleRate:
+        analyticsService.instance.settings.cdnSettings
+          .autoInstrumentationSettings?.sampleRate ?? 0,
     })
 
     const sandbox = new Sandbox(