prometheus cheatsheet

Every export of @jennifer/prometheus. The guide is the introduction.

jennifer

import "@jennifer/prometheus/" as prometheus;

Functions

Function	Returns	Does
`prometheus.counter(name as string, help as string)`	`Metric`	Build an empty counter metric
`prometheus.gauge(name as string, help as string)`	`Metric`	Build an empty gauge metric
`prometheus.histogram(name as string, help as string, buckets as list of float)`	`Metric`	Build an empty histogram metric
`prometheus.summary(name as string, help as string, quantiles as list of float)`	`Metric`	Build an empty summary metric
`prometheus.observe(metric as Metric, labels as map of string to string, value as float)`	`Metric`	Record an observation for a label set, returning a new Metric (value-semantic)
`prometheus.observeAt(metric as Metric, labels as map of string to string, value as float, timestampMs as int)`	`Metric`	Record an observation carrying an explicit millisecond timestamp, returning a new Metric (value-semantic)
`prometheus.render(metrics as list of Metric)`	`string`	Render a list of metrics as the Prometheus text exposition format
`prometheus.pushgatewayPath(job as string, grouping as map of string to string)`	`string`	Build the Pushgateway URL path for a job and a set of grouping labels
`prometheus.query(base as string, promql as string)`	`Result`	Run an instant query against `base` (a Prometheus server URL) via `/api/v1/query`, returning the parsed result set
`prometheus.queryRange(base as string, promql as string, start as string, end as string, step as string)`	`Result`	Run a range query against `base` via `/api/v1/query_range`, returning the parsed result matrix

One recorded series of a metric, keyed by its label set.

Field	Meaning
`labels as map of string to string`	the label name/value pairs ({} for none)
`value as float`	the sample value (Counter / Gauge)
`count as float`	the observation count (Histogram / Summary)
`sum as float`	the observation sum (Histogram / Summary)
`buckets as list of float`	cumulative per-bucket counts (Histogram; parallel to Metric.buckets)
`observations as list of float`	the raw observed values (Summary; source for quantiles)
`timestamp as int`	an explicit millisecond timestamp, appended after the value when set
`hasTimestamp as bool`	whether `timestamp` is present and should be rendered

A metric family.

Field	Meaning
`name as string`	the metric name (`[a-zA-Z_:][a-zA-Z0-9_:]*`)
`help as string`	the HELP text (empty to omit the HELP line)
`type as MetricType`	the metric type
`samples as list of Sample`	the recorded series
`buckets as list of float`	the histogram upper bounds (`le`), ascending; empty for other types
`quantiles as list of float`	the summary quantiles in `[0, 1]`, ascending; empty for other types

One (timestamp, value) point of a query result series.

Field	Meaning
`timestamp as float`	the sample time as a Unix timestamp (seconds)
`value as float`	the sample value

One result series.

Field	Meaning
`metric as map of string to string`	the series label set
`values as list of Point`	the series points

A parsed query result set.

Field	Meaning
`resultType as string`	"vector", "matrix", "scalar", or "string" (a string result whose text is not a number throws)
`series as list of Series`	the result series

A Prometheus metric type.