What causes "LLM stream request timed out" in flipped-aurora/gin-vue-admin?

Increase the timeout value passed to streamLLMRequest/llmAutoSSEStream for long generations

flipped-aurora/gin-vue-admin · error · Error

LLM stream request timed out

Error message

LLM stream request timed out

What it means

streamLLMRequest wraps the whole stream consumption in a timeout via AbortController. If the timeout fires and the fetch is aborted, the resulting AbortError is replaced with a clearer 'LLM stream request timed out' error.

Source

Thrown at web/src/api/autoCode.js:234

    if (buffer.trim()) {
      const trimmed = buffer.trim()
      if (trimmed.startsWith('data:')) {
        const dataStr = trimmed.slice(5).trim()
        if (dataStr && dataStr !== '[DONE]') {
          try {
            const event = JSON.parse(dataStr)
            handleSSEEvent(event)
          } catch {
            state.answerText += dataStr
          }
        }
      }
    }

    return buildLLMStreamResult(state)
  } catch (error) {
    if (controller.signal.aborted && timeoutTriggered) {
      throw new Error('LLM stream request timed out')
    }
    throw error
  } finally {
    if (timeoutId) timerHost.clearTimeout(timeoutId)
  }
}

// 通用 SSE 流式请求导出
export const llmAutoSSEStream = (data, options = {}) =>
  streamLLMRequest(LLM_AUTO_SSE_URL, data, options)

// picture 页专用：预设 mode 为 newCreateWeb
export const createWebStream = (data, options = {}) =>
  llmAutoSSEStream({ mode: 'newCreateWeb', ...data }, options)

export const preview = (data) => {
  return service({
    url: '/autoCode/preview',

View on GitHub (pinned to 3136500ef3)

Solutions

Increase the timeout value passed to streamLLMRequest/llmAutoSSEStream for long generations
Check upstream LLM provider latency/health
Verify the backend proxy streams chunks promptly instead of buffering until completion
Implement resumable/retry logic so partial output is preserved

Example fix

// before
streamLLMRequest(url, { prompt })
// after
streamLLMRequest(url, { prompt, timeout: 300000 }) // 5min for long generations

Defensive patterns

Strategy: retry

Validate before calling

const TIMEOUT_MS = 300000
if ((options.timeout ?? TIMEOUT_MS) < 60000) {
  console.warn('Timeout below 60s may abort long LLM generations')
}

Type guard

function isTimeoutError(err, timeoutFlagRef) {
  return timeoutFlagRef.timeoutTriggered && err.name === 'AbortError'
}

Try / catch

try {
  return await streamLLMRequest(url, { ...options, timeout: 300000 })
} catch (e) {
  if (e.message === 'LLM stream request timed out') {
    return retryWithBackoff(() => streamLLMRequest(url, options), 2)
  }
  throw e
}

Prevention

Set timeouts relative to expected generation length (long code output needs minutes)
Surface partial state.answerText to the user on timeout instead of discarding it
Monitor upstream LLM provider latency

When it happens

Trigger: The LLM stream takes longer than the configured timeout (options.timeout / default) with no completion; fetch rejects with AbortError while timeoutTriggered is true.

Common situations: LLM provider is slow or hung generating a long answer; upstream network stall; timeout set too low for long code-generation responses.

Understand the failure class

Timeouts: ETIMEDOUT, deadlines, and hung requests — what actually expires when a request times out.

Related errors

AI-assisted analysis of flipped-aurora/gin-vue-admin@3136500ef3 (2026-08-31). Data as JSON: /api/errors/f6f834ab184e1099. Report an issue: GitHub.