Knowledge MaxxingMetric
A metric is a numeric measurement tracked over time, such as requests per second or error rate, used to understand a system's overall behavior and health.
Prerequisite: Log
What matters
Because metrics are aggregated numbers, they're well suited to spotting trends and setting alerts, though they don't show the specific detail of any one individual event the way a log does.
Example
A dashboard might track the metric 'average response time' every minute, revealing a gradual slowdown trend that a single log entry wouldn't show on its own.
Common misconception
A metric spiking does not tell you exactly why something went wrong; metrics reveal that something changed, while logs or traces are needed to find the specific cause.
Recall question
Why is a metric well suited to spotting a trend but not to explaining a specific cause?