Ishga tayyor · Logging, metrics va tracing

Daraja: Ilg'or

Productiondagi muammoni taxmin emas, signal va correlation orqali topish.

Tushuntirish

## Ishdagi vazifa Log nima bo‘lganini, metric tizim holatini, trace esa request servislar bo‘ylab qayerda vaqt yo‘qotganini ko‘rsatadi. Strukturali logda timestamp, level, service, trace ID va xavfsiz kontekst bo‘ladi. Parol, token, to‘liq karta yoki shaxsiy ma’lumot loglanmaydi.

## Signal Request rate, error rate va duration asosiy golden signallardir; DB pool, queue lag va JVM memory ham kuzatiladi. Alert foydalanuvchiga ta’sir qiladigan holatga bog‘lanadi, har bitta xatoga emas. Health endpoint dependency holatini ko‘rsatadi, lekin business flow ishlashini yakka o‘zi isbotlamaydi.

## Manba [OpenTelemetry Java](https://opentelemetry.io/docs/languages/java/)

Kod misoli

log.info("order_created orderId={} itemCount={} traceId={}",
         order.id(), order.items().size(), traceId);
// Counter: orders.created; Timer: http.server.duration

Keng tarqalgan xatolar

• Exceptionni context va stack tracesiz faqat `Xato bo‘ldi` deb yozish.

Mashqlar

• Bitta requestni frontenddan DBgacha correlation ID bilan kuzating.
• Latency p95 va 5xx rate uchun foydali alert sharti yozing.

Teglar: observability, ishga-tayyor, production, career-04

Interaktiv ko'rinish (viktorina, izohlar, tarix)