OpenTelemetry 익스포트

Seizn 트레이스를 선호하는 관측성 플랫폼으로 익스포트하여 통합 모니터링을 구현하세요.

0160초 개요

Seizn은 모든 검색 작업에 대해 상세 트레이스를 생성합니다. OpenTelemetry 익스포터를 설정하면 이러한 트레이스가 자동으로 선호하는 백엔드로 전송됩니다.

✓Zero-config OTLP export with automatic span creation
✓Rich attributes including latency, token counts, and costs
✓Context propagation for distributed tracing

Supported Backends

DatadogGrafana TempoJaegerHoneycombNew RelicZipkin

02Setup

Configure OTLP export via environment variables.

Environment Variablesbash

# Enable OTLP export in your Seizn client
# Set these environment variables

# Required: OTLP endpoint
export SEIZN_OTLP_ENDPOINT="https://your-collector:4318"

# Optional: Authentication
export SEIZN_OTLP_HEADERS="Authorization=Bearer your-token"

# Optional: Service name (default: seizn-client)
export SEIZN_SERVICE_NAME="my-rag-app"

03설정 예제

Datadog

Seizn 트레이스를 Datadog APM으로 전송합니다.

Datadog Configurationbash

# Option 1: Direct to Datadog (via Agent)
export SEIZN_OTLP_ENDPOINT="http://localhost:4318"
# Datadog Agent must have OTLP receiver enabled

# Option 2: Direct to Datadog API
export SEIZN_OTLP_ENDPOINT="https://trace.agent.datadoghq.com"
export SEIZN_OTLP_HEADERS="DD-API-KEY=your-datadog-api-key"

# Enable in Seizn
export SEIZN_TELEMETRY_ENABLED="true"

datadog-agent.yamlyaml

otlp_config:
  receiver:
    protocols:
      http:
        endpoint: 0.0.0.0:4318

Grafana Tempo

Seizn 트레이스를 Grafana Tempo로 전송합니다.

Grafana Cloud Configurationbash

# Grafana Cloud
export SEIZN_OTLP_ENDPOINT="https://tempo-us-central1.grafana.net/tempo"
export SEIZN_OTLP_HEADERS="Authorization=Basic $(echo -n 'instance-id:api-token' | base64)"

# Self-hosted Tempo
export SEIZN_OTLP_ENDPOINT="http://tempo:4318"

# Enable in Seizn
export SEIZN_TELEMETRY_ENABLED="true"

Jaeger

Seizn 트레이스를 Jaeger로 전송합니다.

Jaeger Configurationbash

# Jaeger with OTLP receiver (v1.35+)
export SEIZN_OTLP_ENDPOINT="http://jaeger:4318"
export SEIZN_TELEMETRY_ENABLED="true"

Docker Compose Examplebash

# Run Jaeger with OTLP support
docker run -d --name jaeger \
  -e COLLECTOR_OTLP_ENABLED=true \
  -p 16686:16686 \
  -p 4318:4318 \
  jaegertracing/all-in-one:latest

04Production Tips

Sampling Strategy

Control trace volume with head-based or tail-based sampling.

typescript

import { Seizn } from 'seizn';

const client = new Seizn({
  apiKey: process.env.SEIZN_API_KEY,
  telemetry: {
    enabled: true,
    samplingRate: 0.1, // Sample 10% of traces
    // Or use tail-based sampling
    sampleOnlyErrors: true,
    sampleSlowRequests: {
      enabled: true,
      thresholdMs: 1000,
    },
  },
});

Custom Attributes

Add resource and span attributes for better filtering.

typescript

const client = new Seizn({
  apiKey: process.env.SEIZN_API_KEY,
  telemetry: {
    enabled: true,
    resourceAttributes: {
      'deployment.environment': 'production',
      'service.version': '1.2.3',
      'service.namespace': 'rag-apps',
    },
    spanAttributes: {
      'user.tier': 'enterprise',
      'feature.flag.rerank': 'enabled',
    },
  },
});

Batch Configuration

Optimize export performance with batch settings.

typescript

const client = new Seizn({
  apiKey: process.env.SEIZN_API_KEY,
  telemetry: {
    enabled: true,
    batchConfig: {
      maxQueueSize: 2048,
      scheduledDelayMs: 5000,
      maxExportBatchSize: 512,
    },
  },
});

05문제 해결

Issue	Cause	Solution
No traces appearing	Telemetry disabled or endpoint wrong	Set SEIZN_TELEMETRY_ENABLED=true and verify endpoint
Connection refused	Collector not running or wrong port	Check collector status and port 4318
Missing spans in trace	Sampling rate too low	Increase sampling rate or disable sampling

Debug Mode

Enable debug logging to troubleshoot telemetry issues.

bash

export SEIZN_TELEMETRY_DEBUG="true"
export OTEL_LOG_LEVEL="debug"

LlamaIndex Integration Back to Integrations