chore(web): Server side search telemetry (#652)

brendan-kellam · web-flow · commit 76dc2f5a1205 · 2025-12-03T16:04:36.000-08:00
diff --git a/.env.development b/.env.development
@@ -21,7 +21,7 @@ AUTH_URL="http://localhost:3000"
 
 DATA_CACHE_DIR=${PWD}/.sourcebot  # Path to the sourcebot cache dir (ex. ~/sourcebot/.sourcebot)
 SOURCEBOT_PUBLIC_KEY_PATH=${PWD}/public.pem
-# CONFIG_PATH=${PWD}/config.json # Path to the sourcebot config file (if one exists)
+CONFIG_PATH=${PWD}/config.json # Path to the sourcebot config file (if one exists)
 
 # Email
 # EMAIL_FROM_ADDRESS="" # The from address for transactional emails.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -7,9 +7,12 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+<!-- @NOTE: release new MCP version on package release! -->
+
 ### Added
 - Added `ALWAYS_INDEX_FILE_PATTERNS` environment variable to allow specifying a comma seperated list of glob patterns matching file paths that should always be indexed, regardless of size or # of trigrams. [#631](https://github.com/sourcebot-dev/sourcebot/pull/631)
 - Added button to explore menu to toggle cross-repository search. [#647](https://github.com/sourcebot-dev/sourcebot/pull/647)
+- Added server side telemetry for search metrics. [#652](https://github.com/sourcebot-dev/sourcebot/pull/652)
 
 ### Fixed
 - Fixed issue where single quotes could not be used in search queries. [#629](https://github.com/sourcebot-dev/sourcebot/pull/629)
diff --git a/packages/mcp/src/client.ts b/packages/mcp/src/client.ts
@@ -8,7 +8,6 @@ export const search = async (request: SearchRequest): Promise<SearchResponse | S
         method: 'POST',
         headers: {
             'Content-Type': 'application/json',
-            'X-Org-Domain': '~',
             ...(env.SOURCEBOT_API_KEY ? { 'X-Sourcebot-Api-Key': env.SOURCEBOT_API_KEY } : {})
         },
         body: JSON.stringify(request)
@@ -26,7 +25,6 @@ export const listRepos = async (): Promise<ListRepositoriesResponse | ServiceErr
         method: 'GET',
         headers: {
             'Content-Type': 'application/json',
-            'X-Org-Domain': '~',
             ...(env.SOURCEBOT_API_KEY ? { 'X-Sourcebot-Api-Key': env.SOURCEBOT_API_KEY } : {})
         },
     }).then(response => response.json());
@@ -43,7 +41,6 @@ export const getFileSource = async (request: FileSourceRequest): Promise<FileSou
         method: 'POST',
         headers: {
             'Content-Type': 'application/json',
-            'X-Org-Domain': '~',
             ...(env.SOURCEBOT_API_KEY ? { 'X-Sourcebot-Api-Key': env.SOURCEBOT_API_KEY } : {})
         },
         body: JSON.stringify(request)
diff --git a/packages/mcp/src/index.ts b/packages/mcp/src/index.ts
@@ -76,6 +76,7 @@ server.tool(
             contextLines: env.DEFAULT_CONTEXT_LINES,
             isRegexEnabled: true,
             isCaseSensitivityEnabled: caseSensitive,
+            source: 'mcp'
         });
 
         if (isServiceError(response)) {
diff --git a/packages/mcp/src/schemas.ts b/packages/mcp/src/schemas.ts
@@ -31,6 +31,7 @@ export const searchOptionsSchema = z.object({
 
 export const searchRequestSchema = z.object({
     query: z.string(),                                // The zoekt query to execute.
+    source: z.string().optional(),                    // The source of the search request.
     ...searchOptionsSchema.shape,
 });
 
diff --git a/packages/web/package.json b/packages/web/package.json
@@ -156,6 +156,7 @@
     "openai": "^4.98.0",
     "parse-diff": "^0.11.1",
     "posthog-js": "^1.161.5",
+    "posthog-node": "^5.15.0",
     "pretty-bytes": "^6.1.1",
     "psl": "^1.15.0",
     "react": "^19.2.1",
diff --git a/packages/web/src/app/[domain]/components/searchBar/useSuggestionsData.ts b/packages/web/src/app/[domain]/components/searchBar/useSuggestionsData.ts
@@ -55,6 +55,7 @@ export const useSuggestionsData = ({
             query: `file:${suggestionQuery}`,
             matches: 15,
             contextLines: 1,
+            source: 'search-bar-file-suggestions'
         }),
         select: (data): Suggestion[] => {
             if (isServiceError(data)) {
@@ -75,6 +76,7 @@ export const useSuggestionsData = ({
             query: `sym:${suggestionQuery.length > 0 ? suggestionQuery : ".*"}`,
             matches: 15,
             contextLines: 1,
+            source: 'search-bar-symbol-suggestions'
         }),
         select: (data): Suggestion[] => {
             if (isServiceError(data)) {
diff --git a/packages/web/src/app/[domain]/search/useStreamedSearch.ts b/packages/web/src/app/[domain]/search/useStreamedSearch.ts
@@ -129,7 +129,8 @@ export const useStreamedSearch = ({ query, matches, contextLines, whole, isRegex
                         whole,
                         isRegexEnabled,
                         isCaseSensitivityEnabled,
-                    }),
+                        source: 'sourcebot-web-client'
+                    } satisfies SearchRequest),
                     signal: abortControllerRef.current.signal,
                 });
 
diff --git a/packages/web/src/app/api/(server)/search/route.ts b/packages/web/src/app/api/(server)/search/route.ts
@@ -4,6 +4,7 @@ import { search, searchRequestSchema } from "@/features/search";
 import { isServiceError } from "@/lib/utils";
 import { NextRequest } from "next/server";
 import { schemaValidationError, serviceErrorResponse } from "@/lib/serviceError";
+import { captureEvent } from "@/lib/posthog";
 
 export const POST = async (request: NextRequest) => {
     const body = await request.json();
@@ -16,8 +17,14 @@ export const POST = async (request: NextRequest) => {
 
     const {
         query,
+        source,
         ...options
     } = parsed.data;
+
+    await captureEvent('api_code_search_request', {
+        source: source ?? 'unknown',
+        type: 'blocking',
+    });
     
     const response = await search({
         queryType: 'string',
@@ -28,5 +35,6 @@ export const POST = async (request: NextRequest) => {
     if (isServiceError(response)) {
         return serviceErrorResponse(response);
     }
+
     return Response.json(response);
 }
diff --git a/packages/web/src/app/api/(server)/stream_search/route.ts b/packages/web/src/app/api/(server)/stream_search/route.ts
@@ -1,6 +1,7 @@
 'use server';
 
 import { streamSearch, searchRequestSchema } from '@/features/search';
+import { captureEvent } from '@/lib/posthog';
 import { schemaValidationError, serviceErrorResponse } from '@/lib/serviceError';
 import { isServiceError } from '@/lib/utils';
 import { NextRequest } from 'next/server';
@@ -15,9 +16,15 @@ export const POST = async (request: NextRequest) => {
 
     const {
         query,
+        source,
         ...options
     } = parsed.data;
 
+    await captureEvent('api_code_search_request', {
+        source: source ?? 'unknown',
+        type: 'streamed',
+    });
+
     const stream = await streamSearch({
         queryType: 'string',
         query,
diff --git a/packages/web/src/features/chat/components/chatBox/useSuggestionsData.ts b/packages/web/src/features/chat/components/chatBox/useSuggestionsData.ts
@@ -47,6 +47,7 @@ export const useSuggestionsData = ({
                 query: query.join(' '),
                 matches: 10,
                 contextLines: 1,
+                source: 'chat-file-suggestions'
             }))
         },
         select: (data): FileSuggestion[] => {
diff --git a/packages/web/src/features/search/searchApi.ts b/packages/web/src/features/search/searchApi.ts
@@ -2,13 +2,12 @@ import { sew } from "@/actions";
 import { getRepoPermissionFilterForUser } from "@/prisma";
 import { withOptionalAuthV2 } from "@/withAuthV2";
 import { PrismaClient, UserWithAccounts } from "@sourcebot/db";
-import { createLogger, env, hasEntitlement } from "@sourcebot/shared";
+import { env, hasEntitlement } from "@sourcebot/shared";
 import { QueryIR } from './ir';
 import { parseQuerySyntaxIntoIR } from './parser';
 import { SearchOptions } from "./types";
 import { createZoektSearchRequest, zoektSearch, zoektStreamSearch } from './zoektSearcher';
 
-const logger = createLogger("searchApi");
 
 type QueryStringSearchRequest = {
     queryType: 'string';
diff --git a/packages/web/src/features/search/types.ts b/packages/web/src/features/search/types.ts
@@ -94,6 +94,7 @@ export type SearchOptions = z.infer<typeof searchOptionsSchema>;
 
 export const searchRequestSchema = z.object({
     query: z.string(),                                // The zoekt query to execute.
+    source: z.string().optional(),                    // The source of the search request.
     ...searchOptionsSchema.shape,
 });
 export type SearchRequest = z.infer<typeof searchRequestSchema>;
diff --git a/packages/web/src/lib/posthog.ts b/packages/web/src/lib/posthog.ts
@@ -0,0 +1,65 @@
+import { PostHog } from 'posthog-node'
+import { env } from '@sourcebot/shared'
+import { RequestCookies } from 'next/dist/compiled/@edge-runtime/cookies';
+import * as Sentry from "@sentry/nextjs";
+import { PosthogEvent, PosthogEventMap } from './posthogEvents';
+import { cookies } from 'next/headers';
+
+/**
+ * @note: This is a subset of the properties stored in the
+ * ph_phc_<id>_posthog cookie.
+ */
+export type PostHogCookie = {
+    distinct_id: string;
+}
+
+const isPostHogCookie = (cookie: unknown): cookie is PostHogCookie => {
+    return typeof cookie === 'object' &&
+        cookie !== null &&
+        'distinct_id' in cookie;
+}
+
+/**
+* Attempts to retrieve the PostHog cookie from the given cookie store, returning
+* undefined if the cookie is not found or is invalid.
+*/
+const getPostHogCookie = (cookieStore: Pick<RequestCookies, 'get'>): PostHogCookie | undefined => {
+    const phCookieKey = `ph_${env.POSTHOG_PAPIK}_posthog`;
+    const cookie = cookieStore.get(phCookieKey);
+
+    if (!cookie) {
+        return undefined;
+    }
+
+    const parsedCookie = (() => {
+        try {
+            return JSON.parse(cookie.value);
+        } catch (e) {
+            Sentry.captureException(e);
+            return null;
+        }
+    })();
+
+    if (isPostHogCookie(parsedCookie)) {
+        return parsedCookie;
+    }
+
+    return undefined;
+}
+
+export async function captureEvent<E extends PosthogEvent>(event: E, properties: PosthogEventMap[E]) {
+    const cookieStore = await cookies();
+    const cookie = getPostHogCookie(cookieStore);
+
+    const posthog = new PostHog(env.POSTHOG_PAPIK, {
+        host: 'https://us.i.posthog.com',
+        flushAt: 1,
+        flushInterval: 0
+    })
+
+    posthog.capture({
+        event,
+        properties,
+        distinctId: cookie?.distinct_id ?? '',
+    });
+}
diff --git a/packages/web/src/lib/posthogEvents.ts b/packages/web/src/lib/posthogEvents.ts
@@ -313,6 +313,11 @@ export type PosthogEventMap = {
         durationMs: number,
         // Whether or not the user is searching all repositories.
         isGlobalSearchEnabled: boolean,
-    }
+    },
+    //////////////////////////////////////////////////////////////////
+    api_code_search_request: {
+        source: string;
+        type: 'streamed' | 'blocking';
+    },
 } 
 export type PosthogEvent = keyof PosthogEventMap;
diff --git a/yarn.lock b/yarn.lock
@@ -4599,6 +4599,15 @@ __metadata:
   languageName: node
   linkType: hard
 
+"@posthog/core@npm:1.6.0":
+  version: 1.6.0
+  resolution: "@posthog/core@npm:1.6.0"
+  dependencies:
+    cross-spawn: "npm:^7.0.6"
+  checksum: 10c0/28aa907bb21b18587bc5f47c44349ebc834b37d9a4cedb1a18d7b673d4d7cdad2120dda80deceaee707b2f52333e9a08a8e591e1fc4066521ce05e820b76309f
+  languageName: node
+  linkType: hard
+
 "@prisma/client@npm:6.2.1":
   version: 6.2.1
   resolution: "@prisma/client@npm:6.2.1"
@@ -8245,6 +8254,7 @@ __metadata:
     parse-diff: "npm:^0.11.1"
     postcss: "npm:^8"
     posthog-js: "npm:^1.161.5"
+    posthog-node: "npm:^5.15.0"
     pretty-bytes: "npm:^6.1.1"
     psl: "npm:^1.15.0"
     react: "npm:^19.2.1"
@@ -17235,6 +17245,15 @@ __metadata:
   languageName: node
   linkType: hard
 
+"posthog-node@npm:^5.15.0":
+  version: 5.15.0
+  resolution: "posthog-node@npm:5.15.0"
+  dependencies:
+    "@posthog/core": "npm:1.6.0"
+  checksum: 10c0/7db929453cefc9b2d0017e1f7c9ffe7e6ecd2a495ee9861bd5e8f3873f72fa29a318dd7bccf10eb15639e1860aab396d4be502af88afba96ed15ac8b46d57e9d
+  languageName: node
+  linkType: hard
+
 "preact-render-to-string@npm:6.5.11":
   version: 6.5.11
   resolution: "preact-render-to-string@npm:6.5.11"
