Apa iku distributed tracing lan carane correlation ID bisa makarya?

Question

Accepted Answer

**Distributed tracing** ngetutake sawijining request nalika lumaku sawisning layanan, mula sira bisa lihat jalaning end-to-end lan nemoni ngendi latency utawa errors dumadi. **Correlation/trace ID** ngiket kabeh spans saka satunggal request.

## Carane makarya

Layanan pertama nggawe **trace ID**; saben downstream call ngrambah iku (lan parent **span ID**) ing headers.

```http
GET /checkout HTTP/1.1
traceparent: 00-4bf92f3577b34da6a3ce929d0e0e4736-00f067aa0ba902b7-01
# trace-id ───────────────────────────^   ^── span-id
```

```text
Trace abc123:
  Gateway      ├──────────────────────────────┤ 220ms
   Orders        ├────────────┤ 90ms
    Payments        ├──────┤ 60ms   ← slow span found here
    Inventory       ├──┤ 15ms
```

## Apa sing dadi kauntungan

- **Latency breakdown** — layanan/span mana sing alon.
- **Error localization** — hop mana sing gagal.
- **Dependency map** — apa sejatine sing manggil apa.

Peralatan: OpenTelemetry (instrumentation), Jaeger / Zipkin (storage + UI).

## Pitfalls

```text
⚠️ If one service drops the trace header, the trace breaks
⚠️ 100% sampling is expensive — sample, but keep errors
⚠️ Logs are useless across services without the correlation ID attached
```

## Mengapa penting

Tanpa tracing, debugging slow utawa failing request sisining puluhan layanan kuwi mung tebakan; kanthi iku, sira bisa lihat persis hop mana sing tanggung jawab.

Ngrambah correlation ID melalui saben log line lan call iku investasi observability sing paling tinggi leverage-ne ing sistem microservices.