handle more status codes returned from cse-copilot (#55350)

Ebonsignori · web-flow · commit b0c439e6a3c7 · 2025-04-29T20:20:58.000Z
diff --git a/data/ui.yml b/data/ui.yml
@@ -49,13 +49,17 @@ search:
     references: Additional docs
     loading_status_message: Loading Copilot response...
     done_loading_status_message: Done loading Copilot response
-    unable_to_answer: Sorry, I'm unable to answer that question. Check that you selected the correct GitHub version or try a different query.
     copy_answer: Copy answer
     copied_announcement: Copied!
     thumbs_up: This answer was helpful
     thumbs_down: This answer was not helpful
     thumbs_announcement: Thank you for your feedback!
     back_to_search: Back to search
+    responses:
+      unable_to_answer: Sorry, I'm unable to answer that question. Check that you selected the correct GitHub version or try a different question.
+      query_too_large: Sorry, your question is too long. Please try shortening it and asking again.
+      asked_too_many_times: Sorry, you've asked too many questions in a short time period. Please wait a few minutes and try again.
+      invalid_query: Sorry, I'm unable to answer that question. Please try asking a different question.
   failure:
     general_title: There was an error loading search results.
     ai_title: There was an error loading Copilot.
diff --git a/src/fixtures/fixtures/data/ui.yml b/src/fixtures/fixtures/data/ui.yml
@@ -49,13 +49,17 @@ search:
     references: Additional docs
     loading_status_message: Loading Copilot response...
     done_loading_status_message: Done loading Copilot response
-    unable_to_answer: Sorry, I'm unable to answer that question. Check that you selected the correct GitHub version or try a different query.
     copy_answer: Copy answer
     copied_announcement: Copied!
     thumbs_up: This answer was helpful
     thumbs_down: This answer was not helpful
     thumbs_announcement: Thank you for your feedback!
     back_to_search: Back to search
+    responses:
+      unable_to_answer: Sorry, I'm unable to answer that question. Check that you selected the correct GitHub version or try a different question.
+      query_too_large: Sorry, your question is too long. Please try shortening it and asking again.
+      asked_too_many_times: Sorry, you've asked too many questions in a short time period. Please wait a few minutes and try again.
+      invalid_query: Sorry, I'm unable to answer that question. Please try asking a different question.
   failure:
     general_title: There was an error loading search results.
     ai_title: There was an error loading Copilot.
diff --git a/src/search/components/input/AskAIResults.tsx b/src/search/components/input/AskAIResults.tsx
@@ -80,27 +80,30 @@ export function AskAIResults({
 
   const [conversationId, setConversationId] = useState<string>('')
 
-  const handleAICannotAnswer = (passedConversationId?: string) => {
+  const handleAICannotAnswer = (
+    passedConversationId?: string,
+    statusCode = 400,
+    uiMessage = t('search.ai.responses.unable_to_answer'),
+  ) => {
     setInitialLoading(false)
     setResponseLoading(false)
     setAICouldNotAnswer(true)
-    const cannedResponse = t('search.ai.unable_to_answer')
     sendAISearchResultEvent({
       sources: [],
-      message: cannedResponse,
+      message: uiMessage,
       eventGroupId: askAIEventGroupId.current,
       couldNotAnswer: true,
-      status: 400,
+      status: statusCode,
       connectedEventId: passedConversationId || conversationId,
     })
-    setMessage(cannedResponse)
-    setAnnouncement(cannedResponse)
+    setMessage(uiMessage)
+    setAnnouncement(uiMessage)
     setReferences([])
     setItem(
       query,
       {
         query,
-        message: cannedResponse,
+        message: uiMessage,
         sources: [],
         aiCouldNotAnswer: true,
         connectedEventId: passedConversationId || conversationId,
@@ -158,17 +161,44 @@ export function AskAIResults({
       try {
         const response = await executeAISearch(router, version, query, debug)
         if (!response.ok) {
-          console.error(
-            `Failed to fetch search results.\nStatus ${response.status}\n${response.statusText}`,
-          )
-          sendAISearchResultEvent({
-            sources: [],
-            message: '',
-            eventGroupId: askAIEventGroupId.current,
-            couldNotAnswer: false,
-            status: response.status,
-          })
-          return setAISearchError()
+          // If there is JSON and the `upstreamStatus` key, the error is from the upstream sever (CSE)
+          let responseJson
+          try {
+            responseJson = await response.json()
+          } catch (error) {
+            console.error('Failed to parse JSON:', error)
+          }
+          const upstreamStatus = responseJson?.upstreamStatus
+          // If there is no upstream status, the error is either on our end or a 500 from CSE, so we can show the error
+          if (!upstreamStatus) {
+            console.error(
+              `Failed to fetch search results.\nStatus ${response.status}\n${response.statusText}`,
+            )
+            sendAISearchResultEvent({
+              sources: [],
+              message: '',
+              eventGroupId: askAIEventGroupId.current,
+              couldNotAnswer: false,
+              status: response.status,
+            })
+            return setAISearchError()
+            // Query invalid - either sensitive question or spam
+          } else if (upstreamStatus === 400 || upstreamStatus === 422) {
+            return handleAICannotAnswer('', upstreamStatus, t('search.ai.responses.invalid_query'))
+            // Query too large
+          } else if (upstreamStatus === 413) {
+            return handleAICannotAnswer(
+              '',
+              upstreamStatus,
+              t('search.ai.responses.query_too_large'),
+            )
+          } else if (upstreamStatus === 429) {
+            return handleAICannotAnswer(
+              '',
+              upstreamStatus,
+              t('search.ai.responses.asked_too_many_times'),
+            )
+          }
         } else {
           setAISearchError(false)
         }
@@ -211,7 +241,7 @@ export function AskAIResults({
                   return
                 }
               } catch (e) {
-                console.error(
+                console.warn(
                   'Failed to parse JSON:',
                   e,
                   'Line:',
@@ -228,7 +258,7 @@ export function AskAIResults({
                 setConversationId(parsedLine.conversation_id)
               } else if (parsedLine.chunkType === 'NO_CONTENT_SIGNAL') {
                 // Serve canned response. A question that cannot be answered was asked
-                handleAICannotAnswer(conversationIdBuffer)
+                handleAICannotAnswer(conversationIdBuffer, 200)
               } else if (parsedLine.chunkType === 'SOURCES') {
                 if (!isCancelled) {
                   sourcesBuffer = sourcesBuffer.concat(parsedLine.sources)
@@ -242,7 +272,11 @@ export function AskAIResults({
                 }
               } else if (parsedLine.chunkType === 'INPUT_CONTENT_FILTER') {
                 // Serve canned response. A spam question was asked
-                handleAICannotAnswer(conversationIdBuffer)
+                handleAICannotAnswer(
+                  conversationIdBuffer,
+                  200,
+                  t('search.ai.responses.invalid_query'),
+                )
               }
               if (!isCancelled) {
                 setAnnouncement('Copilot Response Loading...')
diff --git a/src/search/lib/ai-search-proxy.ts b/src/search/lib/ai-search-proxy.ts
@@ -4,8 +4,6 @@ import got from 'got'
 import { getHmacWithEpoch } from '@/search/lib/helpers/get-cse-copilot-auth'
 import { getCSECopilotSource } from '@/search/lib/helpers/cse-copilot-docs-versions'
 
-const memoryCache = new Map<string, Buffer>()
-
 export const aiSearchProxy = async (req: Request, res: Response) => {
   const { query, version, language } = req.body
 
@@ -43,15 +41,6 @@ export const aiSearchProxy = async (req: Request, res: Response) => {
   ]
   statsd.increment('ai-search.call', 1, diagnosticTags)
 
-  // TODO: Caching here may cause an issue if the cache grows too large. Additionally, the cache will be inconsistent across pods
-  const cacheKey = `${query}:${version}:${language}`
-  if (memoryCache.has(cacheKey)) {
-    statsd.increment('ai-search.cache_hit', 1, diagnosticTags)
-    res.setHeader('Content-Type', 'application/x-ndjson')
-    res.send(memoryCache.get(cacheKey))
-    return
-  }
-
   const startTime = Date.now()
   let totalChars = 0
 
@@ -84,7 +73,10 @@ export const aiSearchProxy = async (req: Request, res: Response) => {
         const errorMessage = `Upstream server responded with status code ${upstreamResponse.statusCode}`
         console.error(errorMessage)
         statsd.increment('ai-search.stream_response_error', 1, diagnosticTags)
-        res.status(500).json({ errors: [{ message: errorMessage }] })
+        res.status(upstreamResponse.statusCode).json({
+          errors: [{ message: errorMessage }],
+          upstreamStatus: upstreamResponse.statusCode,
+        })
         stream.destroy()
       } else {
         // Set response headers
@@ -101,9 +93,11 @@ export const aiSearchProxy = async (req: Request, res: Response) => {
       console.error('Error streaming from cse-copilot:', error)
 
       if (error?.code === 'ERR_NON_2XX_3XX_RESPONSE') {
-        return res
-          .status(400)
-          .json({ errors: [{ message: 'Sorry I am unable to answer this question.' }] })
+        const upstreamStatus = error?.response?.statusCode || 500
+        return res.status(upstreamStatus).json({
+          errors: [{ message: 'Upstream server error' }],
+          upstreamStatus,
+        })
       }
 
       statsd.increment('ai-search.stream_error', 1, diagnosticTags)
