Distribution contains summary statistics for a population of values. It optionally contains a histogram representing the distribution of those values across a set of buckets.
The summary statistics are the count, mean, sum of the squared deviation from the mean, the minimum, and the maximum of the set of population of values. The histogram is based on a sequence of buckets and gives a count of values that fall into each bucket. The boundaries of the buckets are given either explicitly or by formulas for buckets of fixed or exponentially increasing widths.
Although it is not forbidden, it is generally a bad idea to include non-finite values (infinities or NaNs) in the population of values, as this will render the
sum_of_squared_deviation fields meaningless.
Exemplars are example points that may be used to annotate aggregated distribution values. They are metadata that gives information about a particular value added to a Distribution bucket, such as a trace ID that was active when a value was added. They may contain further information, such as a example values and timestamps, origin, etc.
Value of the exemplar point. This value determines to which bucket the exemplar belongs.
The observation (sampling) time of the above value.
Contextual information about the example value. Examples are:
Literal string: type.googleapis.com/google.protobuf.StringValue
Labels dropped during aggregation: type.googleapis.com/google.monitoring.v3.DroppedLabels
There may be only a single attachment of any given message type in a single exemplar, and this is enforced by the system.