1. tts.delivery.succeeded or tts.delivery.failed (with reason/error code)

Fix Plan

To address the issue, we will implement explicit logging and telemetry for the WebChat TTS pipeline. The fix involves adding structured logs for each stage of the pipeline.

Step-by-Step Solution:

1. Add logging for TTS generation:
   • Log tts.generate.started when TTS generation is initiated.
   • Log tts.generate.succeeded with provider/model/audio metadata upon successful generation.
2. Implement routing logging:
   • Log tts.route.selected with the chosen routing method (e.g., webchat inline playback, media attachment, local speaker).
3. Add delivery logging:
   • Log tts.delivery.succeeded upon successful delivery.
   • Log tts.delivery.failed with reason/error code upon delivery failure.
4. Correlate events with a unique ID:
   • Use a correlation ID to link tool invocation to channel delivery events.

Example Code (Node.js):

// Assuming a logger function and a correlation ID generator
function logEvent(event, data) {
  console.log(`${event}: ${JSON.stringify(data)}`);
}

function generateTTS(text) {
  const correlationId = generateCorrelationId();
  logEvent('tts.generate.started', { correlationId });
  
  // TTS generation logic here
  const ttsData = generateTTSData(text);
  logEvent('tts.generate.succeeded', { correlationId, ...ttsData });
  
  return ttsData;
}

function routeTTS(ttsData) {
  const correlationId = ttsData.correlationId;
  const routeMethod = selectRouteMethod(); // Logic to select route method
  logEvent('tts.route.selected', { correlationId, routeMethod });
  
  // Routing logic here
  return routeMethod;
}

function deliverTTS(ttsData, routeMethod) {
  const correlationId = ttsData.correlationId;
  try {
    // Delivery logic here
    logEvent('tts.delivery.succeeded', { correlationId });
  } catch (error) {
    logEvent('tts.delivery.failed', { correlationId, reason: error.message });
  }
}

Verification

To verify the fix, check the gateway logs for the added events (tts.generate.started, tts.generate.succeeded, tts.route.selected, tts.delivery.succeeded, tts.delivery.failed) with their respective data. Ensure that each WebChat TTS trigger emits these events and that failures include actionable reason codes/messages.

Extra Tips

• Ensure proper error handling and logging in each stage of the pipeline.
• Consider surfacing the final TTS delivery status to the user in the WebChat UI for a better user experience.
• Regularly review logs to identify and address common failure points in the TTS pipeline.