metric time series

metric time series


A set of metric data points that all measure the same metrics and share the same dimension field-value pairs.

For example, the following three metric data points form a metric time series. They offer measurements for the same metrics and have the same dimension field-value pairs.

_time=08-05-2019 16:26:42.025 -0700, metric_name:max.size.kb=500, metric_name:current.size.kb=300, group=queue, name=azd
_time=08-05-2019 16:26:41.055 -0700, metric_name:max.size.kb=345, metric_name:current.size.kb=245, group=queue, name=azd
_time=08-05-2019 16:26:40.023 -0700, metric_name:max.size.kb=334, metric_name:current.size.kb=124, group=queue, name=azd

Meanwhile the following metric data point would not be part of that time series because it has a different value for the name dimension.

_time=08-05-2019 16:26:44.013 -0700, metric_name:max.size.kb=421, metric_name:current.size.kb=273, group=queue, name=nzt

You can group metric data points by time series for the purpose of aggregating functions across those time series, such as averages, sums, or rates (in the case of counter metrics).

For more information

In the Metrics manual:

In the Search Reference: