Basic Usage

2019-08-11

Source: vignettes/basic-usage.Rmd

Introduction

Prometheus is an open-source software project recording real-time metrics using HTTP pull model 1. Prometheus records metrics in a time series database and allows for high-dimensionality.

Queries

The promR package offers Promethous class that provides two methods of sourcing data via instant and range query. The package also facilitates sourcing of the relevant metadata. With use of the sample host, the Prometheus class can be created in the following manner.

# library("promR")
prom <- Prometheus$new(host = "http://demo.robustperception.io", port = 9090)

Instant

The instant query sources the data for specific timestamp. The timestamp should be of class returned by as.POSIXct.

# Timestamps can be in rfc3339 or unix format
metrics_instant <- prom$query(query = "go_goroutines", 
                              time = as.POSIXct("2019-08-11 12:25:57 BST"))
name instance job timestamp value port
go_goroutines demo.robustperception.io prometheus 1565522757 295 9090
go_goroutines demo.robustperception.io pushgateway 1565522757 122 9091
go_goroutines demo.robustperception.io alertmanager 1565522757 36 9093
go_goroutines demo.robustperception.io node 1565522757 9 9100

Range query

A range query permits sourcing of the data for a specific period according to a designated interval.

metrics_range <- prom$rangeQuery(
  query = "go_goroutines",
  start = as.numeric(as.POSIXct("2019-08-11 12:20:32 BST")),
  end = as.numeric(as.POSIXct("2019-08-11 12:30:46 BST")),
  step = "10s"
)
name instance job port timestamp value
go_goroutines demo.robustperception.io prometheus 9090 1565522432 301
go_goroutines demo.robustperception.io prometheus 9090 1565522442 301
go_goroutines demo.robustperception.io prometheus 9090 1565522452 301
go_goroutines demo.robustperception.io prometheus 9090 1565522462 301
go_goroutines demo.robustperception.io prometheus 9090 1565522472 301
go_goroutines demo.robustperception.io prometheus 9090 1565522482 301

Metadata

Metadata can be sourced using the metadataQuery call.

metrics_metadata <- prom$metadataQuery(match_target = '{job="node"}')
knitr::kable(head(metrics_metadata))
metric type help unit instance job
node_netstat_Udp_InErrors unknown Statistic UdpInErrors. demo.robustperception.io:9100 node
node_network_address_assign_type gauge address_assign_type value of /sys/class/net/. demo.robustperception.io:9100 node
node_timex_tick_seconds gauge Seconds between clock ticks. demo.robustperception.io:9100 node
node_memory_SwapTotal_bytes gauge Memory information field SwapTotal_bytes. demo.robustperception.io:9100 node
node_network_dormant gauge dormant value of /sys/class/net/. demo.robustperception.io:9100 node
node_vmstat_pswpin unknown /proc/vmstat information field pswpin. demo.robustperception.io:9100 node

  1. For a detailed overview of the project refer to the project website https://prometheus.io/.↩︎