OpenTelemetry Cheat Sheet

OpenTelemetry Cheat Sheet - BelajarKoding | BelajarKoding

Cheat sheet ini berisi panduan praktis OpenTelemetry (OTel), standar observability open-source untuk traces, metrics, dan logs. Cocok buat developer yang ingin menerapkan distributed tracing di aplikasi Python dan JavaScript dengan backend seperti Jaeger atau Grafana Tempo.

#Konsep Dasar

#Apa Itu OpenTelemetry

OpenTelemetry adalah proyek CNCF yang menyediakan standar, SDK, dan tool untuk menghasilkan dan mengekspor data telemetri (traces, metrics, logs) dari aplikasi. OTel membebaskan kamu dari vendor lock-in karena bisa dikirim ke backend apa pun.

#Tiga Pilar Observability

Pilar	Fungsi	Contoh
Traces	Melacak request melalui multiple service	Request dari API gateway ke database
Metrics	Data numerik agregat	Request count, memory usage, error rate
Logs	Event diskrit dengan timestamp	"User 123 logged in"

#Komponen OpenTelemetry

Komponen

plaintext

Trace: User checkout request
  +-- Span: HTTP GET /checkout (gateway)
       +-- Span: authenticate_user
       +-- Span: get_cart_items
       |    +-- Span: SELECT * FROM cart (database)
       +-- Span: process_payment
            +-- Span: POST /api/payment (external API)

Kind	Arti
`SERVER`	Menerima request (HTTP server, RPC server)
`CLIENT`	Mengirim request (HTTP client, RPC client)
`INTERNAL`	Operasi internal dalam service
`PRODUCER`	Mengirim pesan ke queue
`CONSUMER`	Menerima pesan dari queue

python

# requirements.txt: opentelemetry-api, opentelemetry-sdk,
# opentelemetry-exporter-otlp
 
from opentelemetry import trace
from opentelemetry.sdk.trace import TracerProvider
from opentelemetry.sdk.trace.export import BatchSpanProcessor
from

python

from opentelemetry import trace
 
tracer = trace.get_tracer(__name__)
 
# Cara 1: context manager
with tracer.start_as_current_span("process_order") as span:
    span.set_attribute("order.id",

python

with tracer.start_as_current_span("database_query") as span:
    # Attributes: metadata key-value
    span.set_attribute("db.system", "postgresql")
    span.set_attribute("db.statement", "SELECT * FROM users WHERE id = $1")
    span.set_attribute("db.operation",

python

def handle_request():
    with tracer.start_as_current_span("handle_request") as parent:
        # Span ini otomatis menjadi child dari "handle_request"
        authenticate()
 
        # Span ini juga child dari "handle_request"
        process_data()
 
def authenticate():
    with tracer.start_as_current_span(

javascript

// npm install @opentelemetry/api @opentelemetry/sdk-node
// @opentelemetry/exporter-trace-otlp-grpc
 
const { NodeSDK } = require('@opentelemetry/sdk-node');
const { OTLPTraceExporter } =

bash

# Install instrumentation packages
pip install opentelemetry-distro
opentelemetry-bootstrap -a install
 
# Jalankan aplikasi dengan auto-instrumentation
opentelemetry-instrument \
  --service_name checkout-service \
  --exporter_otlp_endpoint http://localhost:4317 \
  --resource_attributes deployment.environment=production \
  python app.py

Propagator	Header	Keterangan
W3C TraceContext	`traceparent`, `tracestate`	Standar W3C, direkomendasikan
B3 (Zipkin)	`X-B3-TraceId`, `X-B3-SpanId`, dll	Kompatibel dengan Zipkin
Jaeger	`uber-trace-id`	Format Jaeger native

python

from opentelemetry import trace, propagate
from opentelemetry.propagators.composite import CompositePropagator
 
# Inject context ke headers (di sisi client)
headers = {}
propagate.inject(headers)
# headers sekarang berisi traceparent, tracestate, dll
 
# Kirim headers ke service berikutnya
response = requests.post(

python

from opentelemetry.propagate import set_global_textmap
from opentelemetry.trace.propagation.tracecontext import TraceContextTextMapPropagator
from opentelemetry.baggage.propagation import W3CBaggagePropagator
 
# Composite propagator (W3C TraceContext + Baggage)
set_global_textmap(CompositePropagator([
    TraceContextTextMapPropagator(),
    W3CBaggagePropagator(),
]))

python

from opentelemetry import trace, baggage
 
with tracer.start_as_current_span("handle_request") as span:
    # Copy semua baggage ke span attributes untuk visibility
    entries = baggage.get_all()
    for key, value in entries.items():
        span.set_attribute(f"baggage.{key}

python

from opentelemetry import metrics
from opentelemetry.sdk.metrics import MeterProvider
from opentelemetry.sdk.metrics.export import PeriodicExportingMetricReader
from opentelemetry.exporter.otlp.proto.grpc.metric_exporter import OTLPMetricExporter

Tipe	Deskripsi	Contoh
Counter	Nilai yang hanya naik	Total request count
UpDownCounter	Nilai naik atau turun	Active connections
Histogram	Distribusi nilai dalam bucket	Request latency
Observable Counter	Counter dari external source	CPU time
Gauge	Nilai sesaat	Memory usage

python

# Histogram: ukur distribusi latency
request_duration = meter.create_histogram(
    name="http_request_duration_seconds",
    description="HTTP request duration in seconds",
    unit="s"
)
 
# Record latency
request_duration.record(

python

import logging
from opentelemetry import trace
from opentelemetry._logs import set_logger_provider
from opentelemetry.exporter.otlp.proto.grpc._log_exporter import OTLPLogExporter
from

plaintext

App --OTLP--> Collector --OTLP--> Jaeger
                          --OTLP--> Tempo
                          --OTLP--> Prometheus
                          --OTLP--> Datadog

bash

# Jalankan Jaeger all-in-one untuk development
docker run -d --name jaeger \
  -p 16686:16686 \
  -p 4317:4317 \
  -p 4318:4318 \
  jaegertracing/all-in-one:latest
 
# UI tersedia di http://localhost:16686

Tipe Sampling	Cara Kerja
AlwaysOn	Rekam semua trace
AlwaysOff	Jangan rekam trace sama sekali
TraceIDRatioBased	Rekam persentase berdasarkan trace ID
ParentBased	Ikuti keputusan sampler parent (head-based)
OTLP (tail-based)	Sampling di Collector berdasarkan trace lengkap

python

from opentelemetry.sdk.trace.sampling import (
    TraceIdRatioBased, ParentBased, ALWAYS_ON
)
 
sampler = ParentBased(root=TraceIdRatioBased(rate=0.1))  # 10% sampling
provider = TracerProvider(resource=resource, sampler=

API	Interface yang dipakai di kode aplikasi
SDK	Implementasi API yang mengumpulkan dan mengekspor data
Instrumentation Libraries	Auto-instrumentation untuk library populer (Flask, Express, PostgreSQL)
Collector	Service proxy yang menerima, memproses, dan mengekspor telemetri
OTLP	OpenTelemetry Protocol, format standar untuk transfer data

Baca Cheat Sheet Lengkap

#Konsep Dasar

#Apa Itu OpenTelemetry

#Tiga Pilar Observability

#Komponen OpenTelemetry

#Traces dan Spans

#Trace

#Span

#Span Kind

#Manual Instrumentation (Python)

#Setup Dasar

#Membuat Span Manual

#Span Attributes dan Events

#Nested Span

#Manual Instrumentation (JavaScript)

#Setup Dasar

#Membuat Span Manual

#Auto-Instrumentation

#Python Auto-Instrumentation

#JavaScript Auto-Instrumentation

#Context Propagation

#Apa Itu Context Propagation

#Propagator Standar

#W3C TraceContext Header

#Inject dan Extract Manual

#Propagator Configuration

#Baggage

#Apa Itu Baggage

#Menggunakan Baggage

#Baggage Header Format

#Menambahkan Baggage ke Span Attributes

#Metrics

#Membuat dan Menggunakan Metrics

#Tipe Metric

#Histogram dan Observable Gauge

#Logs

#OpenTelemetry Logs

#Collector

#Apa Itu OTel Collector

#Collector Configuration

#Menjalankan Collector dengan Docker

#Backend: Jaeger dan Tempo

#Jaeger

#Grafana Tempo

#Sampling

#Mengapa Sampling

#Konfigurasi Sampler (Python)

#Glossary