5 exercises on observability terms — tracing, metrics, and telemetry standards.
0 / 5 completed
1 / 5
In tracing, what is a span?
A span represents one unit of work in a distributed trace — for example a single HTTP handler, a database call, or an RPC. It carries a name, start and end timestamps (so its duration is known), a set of key-value attributes, status, and events. Spans are nested via parent-child relationships, forming a tree that shows how a request flowed through services. Each span shares a trace ID and has its own span ID. Spans are the building blocks from which a full trace is assembled.
2 / 5
What is a trace?
A trace is the complete, end-to-end record of a single request or transaction as it propagates through a distributed system. It is composed of many spans linked by a shared trace ID and arranged into a parent-child tree (or directed graph). The trace reveals the full call path — which services were involved, in what order, and how long each step took — making it invaluable for diagnosing latency and pinpointing where failures occur. Context propagation (passing the trace ID in headers like traceparent) is what stitches spans across service boundaries.
3 / 5
In observability, what is a metric?
A metric is a numeric measurement captured over time to quantify system behavior. Common types include counters (monotonically increasing totals like request counts), gauges (values that go up and down like memory usage), and histograms (distributions, useful for latency percentiles). Metrics are cheap to store and aggregate, making them ideal for dashboards and alerting on trends. Unlike traces and logs, they are pre-aggregated numbers rather than per-event records, trading detail for efficiency. They are one of the three pillars of observability alongside logs and traces.
4 / 5
What is OpenTelemetry?
OpenTelemetry (OTel) is an open, vendor-neutral observability framework — now a CNCF project — that standardizes how telemetry is generated, collected, and exported. It provides APIs and SDKs for instrumenting code to emit traces, metrics, and logs, plus the OTLP wire protocol and the Collector, a pipeline for receiving, processing, and forwarding telemetry to any backend. Because instrumentation is decoupled from any specific vendor, you can switch observability backends without rewriting application code. OTel has become the de facto standard for telemetry instrumentation.
5 / 5
What are an exemplar and cardinality in metrics?
An exemplar is a representative sample attached to an aggregated metric — typically a trace ID for one of the requests that contributed to a histogram bucket. It bridges metrics and traces: seeing a latency spike on a dashboard, you can jump straight to an example trace behind it. Cardinality is the number of unique combinations of label values on a metric. High cardinality (e.g. labeling by user ID or request ID) explodes the number of time series, dramatically increasing memory and storage cost and potentially overwhelming the monitoring system — so label choices must be bounded.