Which type of statistic shows a count of items over time, like the number of errors?

The type of statistic that shows a count of items over time, such as the number of errors, is counters. Counters are specifically designed to accumulate totals over a specified period, making them ideal for tracking occurrences like errors, transactions, or events. They provide a clear, cumulative count that updates as new data comes in, enabling easy monitoring and historical comparison of different time frames.

In the context of system monitoring or performance evaluation, using counters allows administrators to identify trends, recognize spikes or drops in occurrences, and assess overall system performance or reliability. This can be particularly useful in environments where tracking the number of errors or similar events is critical for maintaining service levels or improving operational efficiency.

Other options do not specifically focus on cumulative counts over time. For example, readings could refer to specific measurements taken at a point in time, while periods-average may involve calculating an average value over a designated time frame rather than counting occurrences. Buckets typically categorize data into groups for analysis instead of tracking individual counts. Therefore, counters are the appropriate choice for tracking counts of items like errors over time.