Swiftpack.co - swift-server-community/SwiftPrometheus as Swift Package

Swiftpack.co is a collection of thousands of indexed Swift packages. Search packages.
See all packages published by swift-server-community.
swift-server-community/SwiftPrometheus 1.0.1
Client side Prometheus library in Swift
⭐️ 105
🕓 32 weeks ago
.package(url: "https://github.com/swift-server-community/SwiftPrometheus.git", from: "1.0.1")

Swift 5.2

SwiftPrometheus, Prometheus client for Swift

A prometheus client for Swift supporting counters, gauges, histograms, summaries and info.


SwiftPrometheus is available through SPM. To include it in your project add the following dependency to your Package.swift:

        .package(url: "https://github.com/swift-server-community/SwiftPrometheus.git", from: "1.0.0")


To see a working demo, see PrometheusExample.

First, we have to create an instance of our PrometheusClient:

import Prometheus
let myProm = PrometheusClient()

Usage with SwiftMetrics

For more details about swift-metrics, please view the GitHub repo.

Starting with SwiftPrometheus 1.0.0-alpha.10 MetricsSystem is no longer directly configured with a PrometheusClient.

Instead, create a PrometheusMetricsFactory instance wrapping a PrometheusClient.

let myProm = PrometheusClient()
MetricsSystem.bootstrap(PrometheusMetricsFactory(client: myProm))

Along with a PrometheusClient, PrometheusMetricsFactory can take a Configuration object setting the following properties:

  • A LabelSanitizer used to sanitize metric names to valid Prometheus values. A default implementation is provided.
  • The Prometheus metric type to use for swift-metrics' Timer. Can be a Histogram or a Summary. Note that when using Histogram, preferredDisplayUnit will not be observed.
  • Default buckets for use by aggregating swift-metrics Recorder instances.

Before Alpha 10

To use SwiftPrometheus with swift-metrics, you need to configure the backend inside the MetricsSystem:

import Metrics
import Prometheus
let myProm = PrometheusClient()

To use prometheus-specific features in a later stage of your program, or to get your metrics out of the system, there is a convenience method added to MetricsSystem:

// This returns the same instance passed in to `.bootstrap()` earlier.
let promInstance = try MetricsSystem.prometheus()

You can then use the same APIs described in the rest of this README.


Counters go up (they can only increase in value), and reset when the process restarts.

let counter = myProm.createCounter(forType: Int.self, named: "my_counter")
counter.inc() // Increment by 1
counter.inc(12) // Increment by given value


Gauges can go up and down, they represent a "point-in-time" snapshot of a value. This is similar to the speedometer of a car.

let gauge = myProm.createGauge(forType: Int.self, named: "my_gauge")
gauge.inc() // Increment by 1
gauge.dec(19) // Decrement by given value
gauge.set(12) // Set to a given value


Histograms track the size and number of events in buckets. This allows for aggregatable calculation of quantiles.

let histogram = myProm.createHistogram(forType: Double.self, named: "my_histogram")
histogram.observe(4.7) // Observe the given value


Summaries track the size and number of events

let summary = myProm.createSummary(forType: Double.self, named: "my_summary")
summary.observe(4.7) // Observe the given value


All metric types support adding labels, allowing for grouping of related metrics. Labels are passed when recording values to your metric as an instance of DimensionLabels, or as an array of (String, String).

Example with a counter:

let counter = myProm.createCounter(forType: Int.self, named: "my_counter", helpText: "Just a counter")

let counter = prom.createCounter(forType: Int.self, named: "my_counter", helpText: "Just a counter")

counter.inc(12, .init([("route", "/users")]))
// OR
counter.inc(12, [("route", "/users")])


Prometheus itself is designed to "pull" metrics from a destination. Following this pattern, SwiftPrometheus is designed to expose metrics, as opposed to submitted/exporting them directly to Prometheus itself. SwiftPrometheus produces a formatted string that Prometheus can parse, which can be integrated into your own application.

By default, this should be accessible on your main serving port, at the /metrics endpoint. An example in Vapor 4 syntax looks like:

app.get("metrics") { req async throws -> String in
    return try await MetricsSystem.prometheus().collect()


Please see SECURITY.md for details on the security process.


All contributions are most welcome!

If you think of some cool new feature that should be included, please create an issue. Or, if you want to implement it yourself, fork this repo and submit a PR!

If you find a bug or have issues, please create an issue explaining your problems. Please include as much information as possible, so it's easier for me to reproduce (Framework, OS, Swift version, terminal output, etc.)


Stars: 105
Last commit: 3 days ago
jonrohan Something's broken? Yell at me @ptrpavlik. Praise and feedback (and money) is also welcome.

Release Notes

32 weeks ago

What's Changed

New Contributors

Full Changelog: https://github.com/swift-server-community/SwiftPrometheus/compare/1.0.0...1.0.1

Swiftpack is being maintained by Petr Pavlik | @ptrpavlik | @swiftpackco | API | Analytics