A metric, essentially, is a numerical representation of data. In the realm of logging and observability, a metric is a numerical value of system data measured over intervals of time. You can use metrics to numerically assess the performance and behaviour of a system over a period of time as counts or measures. Examples of metrics include response times, requests per second, memory usage, processor usage, and uptime.
Metrics help quantify data to derive acceptable performance levels and thresholds. You can then continually compare metrics to pre-established baselines to monitor system performance over time.