sd2k / rocket_prometheus Goto Github PK

View Code? Open in Web Editor NEW

43.0 3.0 11.0 128 KB

Prometheus fairing and handler for Rocket

License: Apache License 2.0

Rust 100.00%

metrics prometheus rocket rust

rocket_prometheus's Introduction

Rocket Prometheus

Prometheus instrumentation for Rocket applications.

Usage

Add this crate to your Cargo.toml alongside Rocket 0.5.0:

[dependencies]
rocket = "0.5.0"
rocket_prometheus = "0.10.0"

Then attach and mount a PrometheusMetrics instance to your Rocket app:

use rocket_prometheus::PrometheusMetrics;

#[rocket::launch]
fn launch() -> _ {
    let prometheus = PrometheusMetrics::new();
    rocket::build()
        .attach(prometheus.clone())
        .mount("/metrics", prometheus)

This will expose metrics like this at the /metrics endpoint of your application:

$ curl localhost:8000/metrics
# HELP rocket_http_requests_duration_seconds HTTP request duration in seconds for all requests
# TYPE rocket_http_requests_duration_seconds histogram
rocket_http_requests_duration_seconds_bucket{endpoint="/metrics",method="GET",status="200",le="0.005"} 2
rocket_http_requests_duration_seconds_bucket{endpoint="/metrics",method="GET",status="200",le="0.01"} 2
rocket_http_requests_duration_seconds_bucket{endpoint="/metrics",method="GET",status="200",le="0.025"} 2
rocket_http_requests_duration_seconds_bucket{endpoint="/metrics",method="GET",status="200",le="0.05"} 2
rocket_http_requests_duration_seconds_bucket{endpoint="/metrics",method="GET",status="200",le="0.1"} 2
rocket_http_requests_duration_seconds_bucket{endpoint="/metrics",method="GET",status="200",le="0.25"} 2
rocket_http_requests_duration_seconds_bucket{endpoint="/metrics",method="GET",status="200",le="0.5"} 2
rocket_http_requests_duration_seconds_bucket{endpoint="/metrics",method="GET",status="200",le="1"} 2
rocket_http_requests_duration_seconds_bucket{endpoint="/metrics",method="GET",status="200",le="2.5"} 2
rocket_http_requests_duration_seconds_bucket{endpoint="/metrics",method="GET",status="200",le="5"} 2
rocket_http_requests_duration_seconds_bucket{endpoint="/metrics",method="GET",status="200",le="10"} 2
rocket_http_requests_duration_seconds_bucket{endpoint="/metrics",method="GET",status="200",le="+Inf"} 2
rocket_http_requests_duration_seconds_sum{endpoint="/metrics",method="GET",status="200"} 0.0011045669999999999
rocket_http_requests_duration_seconds_count{endpoint="/metrics",method="GET",status="200"} 2
# HELP rocket_http_requests_total Total number of HTTP requests
# TYPE rocket_http_requests_total counter
rocket_http_requests_total{endpoint="/metrics",method="GET",status="200"} 2

Metrics

By default this crate tracks two metrics:

rocket_http_requests_total (labels: endpoint, method, status): the total number of HTTP requests handled by Rocket.
rocket_http_requests_duration_seconds (labels: endpoint, method, status): the request duration for all HTTP requests handled by Rocket.

The 'rocket' prefix of these metrics can be changed by setting the ROCKET_PROMETHEUS_NAMESPACE environment variable.

Custom Metrics

Further metrics can be tracked by registering them with the registry of the PrometheusMetrics instance:

#[macro_use]
use once_cell::sync::Lazy;
use rocket::{get, launch, routes};
use rocket_prometheus::{
    prometheus::{opts, IntCounterVec},
    PrometheusMetrics,
};

static NAME_COUNTER: Lazy<IntCounterVec> = Lazy::new(|| {
    IntCounterVec::new(opts!("name_counter", "Count of names"), &["name"])
        .expect("Could not create NAME_COUNTER")
});

#[get("/hello/<name>")]
pub fn hello(name: &str) -> String {
    NAME_COUNTER.with_label_values(&[name]).inc();
    format!("Hello, {}!", name)
}

#[launch]
fn launch() -> _ {
    let prometheus = PrometheusMetrics::new();
    prometheus
        .registry()
        .register(Box::new(NAME_COUNTER.clone()))
        .unwrap();
    rocket::build()
        .attach(prometheus.clone())
        .mount("/", routes![hello])
        .mount("/metrics", prometheus)
}

rocket_prometheus's People

Contributors

Stargazers

Watchers

Forkers

skeen irexiz insertish mbr codeths gimlet2 kirinse dietappy failpark ivashchenkoserhii hansl

rocket_prometheus's Issues

Add a shorthand to use the underlying Prometheus's "default_registry" registry.

 pub fn setup_metrics() -> PrometheusMetrics {
     let default_registry = prometheus::default_registry(); ‣default_registry: &Registry
     PrometheusMetrics::with_registry(default_registry.clone())
 }

Would you be willing to merge this upstream? or create an associated function doing the same?

/metrics endpoint is resolving but no metrics are displayed

I'm using version = "0.10.0-rc.3" and my rocket webserver will resolve /metrics but there is no data returned, even though I have sent several GET AND POST requests to my server.

rocket::build()
.attach(AdHoc::config::())
.manage(pool)
.manage(replier)
.manage(prometheus.clone())
.manage(amqp_pool)
.attach(rocket_governor::LimitHeaderGen::default())
.register("/", catchers!(rocket_governor_catcher)) // catches limit errors and produces a 429 error code
.mount("/ref", routes![refs::get,refs::patch, refs::delete, refs::post, refs::get_all])
.mount("/device", routes![device::get,device::patch, device::delete, device::post, device::get_all])
.mount("/library", routes![library::get,library::patch, library::delete, library::post, library::get_all, library::link_ref_item_to_library, library::get_ref_item_in_library])
.mount("/session", routes![session::get,session::patch, session::delete, session::post, session::get_all])
.mount("/project", routes![project::get,project::patch, project::delete, project::post, project::get_all,project::link_library_to_project, project::get_library_in_project,project::match_detections])
.mount("/metrics", prometheus)
.launch().await?;

Is there a way to exclude /metrics from the reporting

Is there a way to exclude specific routes like metrics to exclude from the prometheus stats?

custom metrics with callback

How to write custom metrics which query data out of a database?

e.g. count of users in database as a gauge

PrometheusMetrics::with_default_registry() panics if called more than once

This is especially a problem when writing rocket-tests, see this minimal working example: https://github.com/henrikhorluck/prom-test

Note that the example uses #4, but I believe the issue is still present in 0.8.0

prepare rocket v0.5

it does not build with current master of rocket ;)

error[E0277]: the trait bound PrometheusMetrics: Fairing is not satisfied
--> src/main.rs:38:17
|
38 | .attach(prometheus.clone())
| ^^^^^^^^^^^^^^^^^^ the trait Fairing is not implemented for PrometheusMetrics

error[E0277]: the trait bound Vec<Route>: From<PrometheusMetrics> is not satisfied
--> src/main.rs:44:10
|
44 | .mount("/metrics", prometheus)
| ^^^^^ the trait From<PrometheusMetrics> is not implemented for Vec<Route>
|
= help: the following implementations were found:
<Vec<T> as From<&[T]>>
<Vec<T> as From<&mut [T]>>
<Vec<T> as From<BinaryHeap<T>>>
<Vec<T> as From<Box<[T]>>>
and 6 others
= note: required because of the requirements on the impl of Into<Vec<Route>> for PrometheusMetrics

sorry, i know this issue is a really early for a unstable rocket version

Update to current version

Hey!
Cool plugin! It works like a charm. Unfortunately tools like Grafana dont support old Prometheus versions as it seems.

The current API specification is prometheus 1.8 I believe. It would be nice if you would implement instant queries (i.e., Prometheus 2.14 specification) such that Grafana works :)

Cheers!

Add intra-doc links to documentation

The various items in the doc comments should use intra-doc links to improve the experience of browsing the docs.