··Observability

Use logs, metrics, traces, and OpenTelemetry to find where requests slow down. — 27 entries

···

Distributed tracing분산 추적

Follow one request across services by connecting each operation’s timing and parent–child relationship.

···

Logs로그

Record individual events with their time and context.

···

Metrics메트릭

Aggregate changing system state into numbers and trends.

···

Span스팬

A tracing unit with an operation’s start, end, attributes, and parent.

···

Trace context propagation추적 문맥 전파

Carry a shared trace ID and parent information across service boundaries.

···

Structured logging구조화 로그

Write logs as stable fields so they can be searched and aggregated.

···

Golden signals황금 신호

Read service health through latency, traffic, errors, and saturation together.

···

Service level objective서비스 수준 목표

Set a target and time window for a user-facing reliability indicator.

···

Error budget오류 예산

Translate an SLO into allowable failures to guide release risk.

···

Tail sampling후단 샘플링

Choose which traces to retain after seeing their completed outcomes.

···

Continuous profiling지속적 프로파일링

Continuously sample function-level CPU use in a running service.

···

OpenTelemetry profilesOpenTelemetry 프로파일 신호

Handle profile data alongside logs, metrics, and traces as a telemetry signal.

···

Trace–profile correlation추적·프로파일 연결

Move from a slow request span to the function costs behind it.

···

GenAI semantic conventions생성 AI 의미 규약

Record model-call attributes under shared field names.

···

Token usage metrics토큰 사용량 메트릭

Aggregate input and output tokens separately for model calls.

···

Agent tool tracing에이전트 도구 추적

Connect model decisions, tool calls, and results in one trace tree.

···

eBPF auto-instrumentationeBPF 자동 계측

Collect request-boundary telemetry without editing application code.

···

CI/CD observabilityCI/CD 관측성

Trace build, test, and deploy stages to locate slowdowns and failures.

···

OTel Arrow pipelineOTel Arrow 파이프라인

Group repeated telemetry attributes by column to improve transport efficiency.

···

Metric cardinality limit메트릭 카디널리티 상한

Cap the number of time series created by label combinations.

···

Native histograms네이티브 히스토그램

Store a distribution in one sample with adaptive buckets.

···

Consistent probability sampling일관 확률 샘플링

Apply a consistent sampling choice to a request across services.

···

Burn-rate alert소진율 경보

Alert on how quickly the error budget is being spent.

···

Exemplars예시 추적 연결

Link an individual observation in an aggregate metric to a trace.

···

Declarative OpenTelemetry configuration선언형 OpenTelemetry 설정

Define instrumentation in a file using a common model across languages.

···

OpenTelemetry Go Logs release candidateOpenTelemetry Go 로그 API 릴리스 후보

The Go Logs API and SDK reached the release candidate stage.

···

Collector pipeline수집기 파이프라인

Receive telemetry, process it, and export it to a destination.

Open as page ↗