In Prometheus, a Gauge is the metric type used to represent a value that can increase and decrease over time, so B is correct. Gauges are suited for “current state” measurements such as current memory usage, number of active sessions, queue depth, temperature, or CPU usage—anything that can move up and down as the system changes.
This contrasts with a Counter (A), which is monotonically increasing (it only goes up, except when a process restarts and the counter resets to zero). Counters are ideal for totals like total HTTP requests served, total errors, or bytes sent, and you typically use rate()/irate() in PromQL to convert counters into per-second rates.
A Summary (C) and Histogram (D) are used for distributions, commonly request latency. Histograms record observations into buckets and can produce percentiles using functions like histogram_quantile(). Summaries compute quantiles on the client side and expose them directly, along with counts and sums. Neither of these is the simplest “single value that goes up and down” type.
In Kubernetes observability, Prometheus is often used to scrape metrics from cluster components (API server, kubelet) and applications. Choosing the right metric type matters operationally: use gauges for instantaneous measurements, counters for event totals, and histograms/summaries for latency distributions. That’s why Prometheus documentation and best practices emphasize understanding metric semantics—because misusing types leads to incorrect alerts and dashboards.
So for a single numeric value that can go up and down, the correct metric type is Gauge, option B.
=========
