Metrics

Metrics are values you can measure over time, such as the number of blocks used on a filesystem, the number of nodes in a cluster, or a temperature reading. Observe ingests this time series data into Event Datasets. Any Event Dataset with a numerical value column can be interpreted as a Metric Dataset in Observe, in Metrics Explorer, a dashboard, a worksheet, or a monitor . Each point in a time series represents a measurement from a single resource, with its name, value, and tags .

A metric has the following major metadata attributes:

• Metric Type: The metric type is used by Observe to interpret the reported values. It is critical to set the Metric Type correctly. For instance, a cumulativeCounter's normal reset could look like a service failure if that metric is mistakenly typed as a gauge or delta.
• Measurement Type: The metric's measurement type describes the data type reported in each point. Observe supports float64 and tdigest as data type.
• Unit: The metrics' unit describes the unit of the measurements, such as kB or kB/s. Observe uses these attributes to display metrics properly in charts. See Units of measurement.