Overview

Overview

Monitor Ideogram AI image generation for latency, credit consumption, and output quality. Key metrics include generation time (typically 5-15 seconds per image depending on model and resolution), credit cost per generation (varies by model version and quality setting), generation success rate, and prompt safety filter rejection rate.

Prerequisites

• Ideogram API account with active credits
• Metrics backend for tracking generation data
• Webhook or polling mechanism for async generation status

Instructions

Step 1: Instrument Image Generation Calls

async function trackedGeneration(prompt: string, options: any) {
  const start = performance.now();
  try {
    const result = await ideogram.generate({ image_request: { prompt, ...options } });
    const duration = performance.now() - start;
    emitHistogram('ideogram_generation_duration_ms', duration, { model: options.model || 'V_2' });
    emitCounter('ideogram_generations_total', 1, { model: options.model || 'V_2', status: 'success' });
    emitCounter('ideogram_credits_used', result.credits_consumed || 1, { model: options.model || 'V_2' });
    return result;
  } catch (err: any) {
    emitCounter('ideogram_generations_total', 1, { status: 'error', reason: err.code || 'unknown' });
    throw err;
  }
}