prometheus.Histogram.observe()

Records a single observation (or measurement) for a histogram metric in the appropriate bucket for the specified label values.

Syntax

Note:

In this syntax diagram, the square brackets ([ ]) are part of the syntax.

observe( value FLOAT, [ labels ] )
  1. value is a value of type FLOAT to observe (for example, a duration or size).
  2. labels is a comma-separated list of label values inside square brackets, one for each label name declared when the metric was created, in the same order. Where the metric declares no label names, use empty brackets [].

Usage

Use the prometheus.Histogram.observe() method to observe a value in the histogram metric.
  • Call once per request/response you want to measure (typically just before sending the response).
  • Use floats; the value is placed into the configured buckets and also contributes to the histogram's sum and count.
This method is typically used for tracking distributions such as request durations or response sizes. The label values change from one call to the next. For example, with the label names ["url", "status"] declared, observe(0.42, ["/orders", "ok"]) records an observation for that URL and status.

Example


-- Define the histogram variable
DEFINE histogram prometheus.Histogram
-- Create a histogram metric
LET histogram = prometheus.Histogram.create("histogram_request_duration",
                                            "request duration in seconds", [0.25,0.5,1],
                                            ["url", "status"])
-- Observe a value
CALL histogram.observe(0.42, ["/orders", "ok"])