instrument-distributed-tracing
정보
이 스킬은 개발자들이 OpenTelemetry를 활용한 분산 추적을 구현하는 데 도움을 줍니다. 자동 및 수동 계측, 컨텍스트 전파, Jaeger와 같은 백엔드 통합을 모두 다룹니다. 지연 문제 디버깅, 마이크로서비스 간 요청 흐름 이해, 레거시 추적 시스템에서 마이그레이션할 때 사용하세요. 이 스킬은 추적 데이터를 로그 및 메트릭과 연계하여 포괄적인 근본 원인 분석을 가능하게 합니다.
빠른 설치
Claude Code
추천npx skills add pjt222/agent-almanac -a claude-code/plugin add https://github.com/pjt222/agent-almanacgit clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/instrument-distributed-tracingClaude Code에서 이 명령을 복사하여 붙여넣어 스킬을 설치하세요
문서
Instrument Distributed Tracing
Implement OpenTelemetry distributed tracing to track requests across microservices and identify performance bottlenecks.
When to Use
- Debugging latency in distributed systems with multiple services
- Understanding request flow and dependencies between microservices
- Identifying slow database queries or external API calls within a transaction
- Correlating traces with logs and metrics for root cause analysis
- Measuring end-to-end latency from user request to response
- Migrating from legacy tracing systems (Zipkin, Jaeger) to OpenTelemetry
- Establishing SLO compliance through detailed latency percentile tracking
Inputs
- Required: List of services to instrument (languages and frameworks)
- Required: Tracing backend choice (Jaeger, Tempo, Zipkin, or vendor SaaS)
- Optional: Existing instrumentation libraries (OpenTracing, Zipkin)
- Optional: Sampling strategy requirements (percentage, rate limiting)
- Optional: Custom span attributes for business-specific metadata
Procedure
See Extended Examples for complete configuration files and templates.
Step 1: Set Up Tracing Backend
Deploy Jaeger or Grafana Tempo to receive and store traces.
Option A: Jaeger all-in-one (development/testing):
# docker-compose.yml
version: '3.8'
services:
jaeger:
image: jaegertracing/all-in-one:1.51
ports:
- "5775:5775/udp" # Zipkin compact thrift
- "6831:6831/udp" # Jaeger compact thrift
- "6832:6832/udp" # Jaeger binary thrift
- "5778:5778" # Serve configs
- "16686:16686" # Jaeger UI
- "14268:14268" # Jaeger HTTP thrift
- "14250:14250" # Jaeger GRPC
- "9411:9411" # Zipkin compatible endpoint
environment:
- COLLECTOR_ZIPKIN_HOST_PORT=:9411
- COLLECTOR_OTLP_ENABLED=true
restart: unless-stopped
Option B: Grafana Tempo (production, scalable):
# docker-compose.yml
version: '3.8'
services:
tempo:
image: grafana/tempo:2.3.0
command: ["-config.file=/etc/tempo.yaml"]
volumes:
- ./tempo.yaml:/etc/tempo.yaml
- tempo-data:/tmp/tempo
ports:
- "3200:3200" # Tempo HTTP
- "4317:4317" # OTLP gRPC
- "4318:4318" # OTLP HTTP
- "9411:9411" # Zipkin
restart: unless-stopped
volumes:
tempo-data:
Tempo configuration (tempo.yaml):
server:
http_listen_port: 3200
distributor:
receivers:
jaeger:
# ... (see EXAMPLES.md for complete configuration)
For production with S3 storage:
storage:
trace:
backend: s3
s3:
bucket: tempo-traces
endpoint: s3.amazonaws.com
region: us-east-1
wal:
path: /tmp/tempo/wal
pool:
max_workers: 100
queue_depth: 10000
Got: Tracing backend accessible, ready to receive traces via OTLP, Jaeger UI or Grafana shows "no traces" initially.
If fail:
- Verify ports not already in use:
netstat -tulpn | grep -E '(4317|16686|3200)' - Check container logs:
docker logs jaegerordocker logs tempo - Test OTLP endpoint:
curl http://localhost:4318/v1/traces -v - For Tempo: validate config syntax with
tempo -config.file=/etc/tempo.yaml -verify-config
Step 2: Instrument Applications (Auto-Instrumentation)
Use OpenTelemetry auto-instrumentation for common frameworks to minimize code changes.
Python with Flask:
pip install opentelemetry-distro opentelemetry-exporter-otlp
opentelemetry-bootstrap -a install
# app.py
from flask import Flask
from opentelemetry import trace
from opentelemetry.sdk.trace import TracerProvider
from opentelemetry.sdk.trace.export import BatchSpanProcessor
from opentelemetry.exporter.otlp.proto.grpc.trace_exporter import OTLPSpanExporter
# ... (see EXAMPLES.md for complete configuration)
Go with Gin framework:
go get go.opentelemetry.io/otel
go get go.opentelemetry.io/otel/exporters/otlp/otlptrace/otlptracegrpc
go get go.opentelemetry.io/otel/sdk/trace
go get go.opentelemetry.io/contrib/instrumentation/github.com/gin-gonic/gin/otelgin
package main
import (
"context"
"github.com/gin-gonic/gin"
"go.opentelemetry.io/otel"
# ... (see EXAMPLES.md for complete configuration)
Node.js with Express:
npm install @opentelemetry/api \
@opentelemetry/sdk-node \
@opentelemetry/auto-instrumentations-node \
@opentelemetry/exporter-trace-otlp-grpc
// tracing.js
const { NodeSDK } = require('@opentelemetry/sdk-node');
const { OTLPTraceExporter } = require('@opentelemetry/exporter-trace-otlp-grpc');
const { getNodeAutoInstrumentations } = require('@opentelemetry/auto-instrumentations-node');
const { Resource } = require('@opentelemetry/resources');
const { SemanticResourceAttributes } = require('@opentelemetry/semantic-conventions');
# ... (see EXAMPLES.md for complete configuration)
Got: Traces from instrumented services appear in Jaeger UI or Grafana, HTTP requests automatically create spans.
If fail:
- Check exporter endpoint is reachable from application
- Verify environment variables:
OTEL_EXPORTER_OTLP_ENDPOINT=http://tempo:4317 - Enable debug logging:
OTEL_LOG_LEVEL=debug(Python),OTEL_LOG_LEVEL=DEBUG(Node.js) - Test with simple span: manually create a span to verify export pipeline
- Check for version conflicts between OpenTelemetry packages
Step 3: Add Manual Instrumentation
Create custom spans for business logic, database queries, and external calls.
Python manual spans:
from opentelemetry import trace
tracer = trace.get_tracer(__name__)
def process_order(order_id):
# Create a span for the entire operation
# ... (see EXAMPLES.md for complete configuration)
Go manual spans:
import (
"context"
"go.opentelemetry.io/otel"
"go.opentelemetry.io/otel/attribute"
"go.opentelemetry.io/otel/codes"
"go.opentelemetry.io/otel/trace"
# ... (see EXAMPLES.md for complete configuration)
Span attributes best practices:
- Use semantic conventions:
http.method,http.status_code,db.system,db.statement - Add business context:
user.id,order.id,product.category - Include resource identifiers:
instance.id,region,availability_zone - Record errors:
span.RecordError(err)andspan.SetStatus(codes.Error, message) - Add events for significant milestones:
span.AddEvent("cache_miss")
Got: Custom spans appear in trace view, parent-child relationships correct, attributes visible in span details, errors highlighted.
If fail:
- Verify context propagation: parent span context passed to child
- Check span names are descriptive and follow naming conventions
- Ensure spans are ended (use
defer span.End()in Go,withblocks in Python) - Review attribute types: strings, ints, bools, floats only
- Validate semantic conventions: use standard attribute names where applicable
Step 4: Implement Context Propagation
Ensure trace context flows across service boundaries and async operations.
HTTP headers propagation (W3C Trace Context):
# Client side (Python with requests)
import requests
from opentelemetry import trace
from opentelemetry.propagate import inject
tracer = trace.get_tracer(__name__)
# ... (see EXAMPLES.md for complete configuration)
// Server side (Go with Gin)
import (
"go.opentelemetry.io/otel"
"go.opentelemetry.io/otel/propagation"
)
# ... (see EXAMPLES.md for complete configuration)
Message queue propagation (Kafka):
# Producer
from opentelemetry.propagate import inject
from kafka import KafkaProducer
producer = KafkaProducer(bootstrap_servers=['kafka:9092'])
# ... (see EXAMPLES.md for complete configuration)
# Consumer
from opentelemetry.propagate import extract
def process_message(msg):
# Extract trace context from Kafka headers
headers = {k: v.decode('utf-8') for k, v in msg.headers}
ctx = extract(headers)
# Continue the trace
with tracer.start_as_current_span("process_order_event", context=ctx):
order_id = json.loads(msg.value)['order_id']
handle_order(order_id)
Async operations (Python asyncio):
import asyncio
from opentelemetry import trace, context
async def async_operation():
# Capture current context
token = context.attach(context.get_current())
try:
with tracer.start_as_current_span("async_database_query"):
await asyncio.sleep(0.1) # Simulated async work
return "result"
finally:
context.detach(token)
Got: Traces span multiple services, trace IDs consistent across service boundaries, parent-child relationships preserved.
If fail:
- Verify W3C Trace Context propagator configured:
otel.propagation.set_global_textmap(TraceContextTextMapPropagator()) - Check headers are passed in HTTP requests
- For Kafka: ensure headers supported by broker version (v0.11+)
- Debug with header inspection: log
traceparentheader value - Use trace visualization to identify broken trace links
Step 5: Configure Sampling Strategies
Implement sampling to reduce trace volume and cost while maintaining visibility.
Sampling strategies:
from opentelemetry.sdk.trace import TracerProvider
from opentelemetry.sdk.trace.sampling import (
ParentBased,
TraceIdRatioBased,
StaticSampler,
Decision
# ... (see EXAMPLES.md for complete configuration)
Tail-based sampling with Tempo:
Configure in tempo.yaml:
overrides:
defaults:
metrics_generator:
processors: [service-graphs, span-metrics]
storage:
path: /tmp/tempo/generator/wal
remote_write:
- url: http://prometheus:9090/api/v1/write
send_exemplars: true
# Tail sampling (requires tempo-query)
ingestion_rate_limit_bytes: 5000000
ingestion_burst_size_bytes: 10000000
Use Grafana Tempo's TraceQL for dynamic sampling:
# Sample traces with errors
{ status = error }
# Sample slow traces (>1s)
{ duration > 1s }
# Sample specific services
{ resource.service.name = "checkout-service" }
Got: Trace volume reduced to target percentage, error traces always sampled, sampling decision visible in trace metadata.
If fail:
- Verify sampler applied before tracer provider initialization
- Check sampling decision attribute in exported spans
- For tail sampling: ensure sufficient buffering (
ingestion_burst_size_bytes) - Monitor dropped traces:
otel_traces_dropped_totalmetric - Test with synthetic high-volume traffic to validate sampling rate
Step 6: Correlate Traces with Metrics and Logs
Link traces to metrics and logs for unified observability.
Add trace IDs to logs (Python):
import logging
from opentelemetry import trace
# Custom log formatter with trace context
class TraceFormatter(logging.Formatter):
def format(self, record):
# ... (see EXAMPLES.md for complete configuration)
Generate metrics from traces (Tempo):
# tempo.yaml
metrics_generator:
registry:
external_labels:
cluster: production
storage:
# ... (see EXAMPLES.md for complete configuration)
This generates Prometheus metrics:
traces_service_graph_request_total- request count between servicestraces_span_metrics_duration_seconds- span duration histogramtraces_spanmetrics_calls_total- span call counts
Query traces from metrics (Grafana):
Add exemplar support to Prometheus datasource in Grafana:
datasources:
- name: Prometheus
type: prometheus
url: http://prometheus:9090
jsonData:
exemplarTraceIdDestinations:
- name: trace_id
datasourceName: Tempo
In Grafana dashboard, enable exemplars:
{
"fieldConfig": {
"defaults": {
"custom": {
"showExemplars": true
}
}
}
}
Got: Clicking metric exemplars opens trace, logs show trace IDs, traces link to logs, unified debugging across signals.
If fail:
- Verify exemplar support enabled in Prometheus (requires v2.26+)
- Check trace ID format matches (32-char hex)
- Ensure metrics generator enabled in Tempo config
- Validate remote write endpoint accessible from Tempo
- Test exemplar queries:
histogram_quantile(0.95, rate(http_request_duration_seconds_bucket[5m])) and on() exemplar
Validation
- Tracing backend receives spans from all instrumented services
- Traces show correct parent-child relationships across services
- Span attributes include semantic conventions and business context
- Context propagates correctly across HTTP calls and message queues
- Sampling strategy reduces trace volume to target percentage
- Error traces always sampled (if using error-aware sampling)
- Trace IDs appear in application logs with correct format
- Grafana shows traces linked from metrics via exemplars
- Log panels have data links to trace viewer
- Trace retention matches configured storage policy
Pitfalls
- Context not propagated: Forgetting to pass
contextto downstream calls breaks traces. Pass context explicitly. - Spans never ended: Missing
defer span.End()(Go) orwithblocks (Python) causes spans to remain open and memory leaks. - Over-instrumentation: Creating spans for every function causes trace bloat. Focus on service boundaries, database calls, and external APIs.
- Missing error recording: Not calling
span.RecordError()loses debugging information. Record errors in spans. - High cardinality attributes: Using unbounded values (user IDs, request bodies) as span attributes causes storage issues. Use sampling or aggregate labels.
- Incorrect span kind: Using wrong span kind (CLIENT vs SERVER vs INTERNAL) affects service graph generation. Follow semantic conventions.
- Sampling before context: Sampling decisions must respect parent trace context. Use
ParentBasedsampler to honor upstream sampling.
Related Skills
correlate-observability-signals- Unified debugging with metrics, logs, and traces linked by trace IDssetup-prometheus-monitoring- Generate metrics from traces using Tempo metrics generatorconfigure-log-aggregation- Add trace IDs to logs for correlation with distributed tracesbuild-grafana-dashboards- Visualize trace-derived metrics and exemplar links in dashboards
GitHub 저장소
연관 스킬
evaluating-llms-harness
테스팅이 Claude Skill은 MMLU, GSM8K를 포함한 60개 이상의 표준화된 학술 과제에서 LLM 성능을 벤치마크하기 위해 lm-evaluation-harness를 실행합니다. 개발자들이 모델 품질을 비교하고, 학습 진행 상황을 추적하거나 학술 결과를 보고할 수 있도록 설계되었습니다. 이 도구는 HuggingFace와 vLLM 모델을 포함한 다양한 백엔드를 지원합니다.
cloudflare-cron-triggers
테스팅이 스킬은 cron 표현식을 사용하여 Worker를 스케줄링하기 위한 Cloudflare Cron Triggers 구현에 관한 포괄적인 지식을 제공합니다. 주기적 작업, 유지보수 작업, 자동화된 워크플로우 설정 방법을 다루며, 잘못된 cron 표현식이나 시간대 문제 같은 일반적인 이슈들을 해결하는 방법을 포함합니다. 개발자들은 이를 통해 스케줄된 핸들러 구성, cron 트리거 테스트, Workflows 및 Green Compute와의 연동 작업을 수행할 수 있습니다.
webapp-testing
테스팅이 Claude Skill은 Python 스크립트를 통해 로컬 웹 애플리케이션을 테스트하기 위한 Playwright 기반 툴킷을 제공합니다. 프론트엔드 검증, UI 디버깅, 스크린샷 캡처, 로그 확인 기능을 지원하며 서버 라이프사이클을 관리합니다. 브라우저 자동화 작업에 사용하되 컨텍스트 오염을 방지하기 위해 소스 코드를 읽지 않고 스크립트를 직접 실행하세요.
finishing-a-development-branch
테스팅이 스킬은 테스트 통과를 확인한 후 체계적인 통합 옵션을 제시하여 개발자가 완성된 작업을 마무리하도록 돕습니다. 구현이 완료된 후 머지, PR 생성, 브랜치 정리와 같은 워크플로우를 안내합니다. 코드가 준비되고 테스트가 완료되었을 때 개발 프로세스를 체계적으로 마무리하기 위해 사용하세요.
