Rust DogStatsd

A DogStatsD client implementation of statsd in Rust.

• Forked from https://github.com/minato128/rust-dogstatsd, which was forked from https://github.com/markstory/rust-statsd
• Adds timing of async functions as well as makes the Client type Cloneable

Using the client library

Add the datadog-statsd package as a dependency in your Cargo.toml file:

[dependencies]
datadog-statsd = "0.2.0"

You need rustc >= 1.31.0 for statsd to work.

You can then get a client instance and start tracking metrics:

// Import the client and config objects.
use datadog_statsd::{Client, ClientConfig};

// Build a config with a prefix and some constant tags
let config = ClientConfig::builder(("127.0.0.1", 8125))
    .prefix("myapp")
    .constant_tags(vec!["common1", "common2:test"])
    .build();

// Make a client from this config
let client = Client::new(&config).unwrap();

Tracking Metrics

Once you've created a client, you can track timers and metrics:

let tags = Some(vec!["tag1", "tag2:test"]);

// Increment a counter by 1
client.incr("some.counter", tags.as_ref());

// Decrement a counter by 1
client.decr("some.counter", tags.as_ref());

// Update a gauge
client.gauge("some.value", 12.0, tags.as_ref());

// Modify a counter by an arbitrary float.
client.count("some.counter", 511.0, tags.as_ref());

// Send a histogram value as a float.
client.histogram("some.histogram", 511.0, tags.as_ref());

Tracking Timers

Timers can be updated using timer(), time(), and time_async():

// Update a timer based on a calculation you've done.
client.timer("operation.duration", 13.4, tags.as_ref());

// Time a closure
client.time("operation.duration", tags.as_ref(), || {
	// Do something expensive.
});

// Time an async closure
let result = client.time_async(
    "operation.duration",
    tags.as_ref(),
    || something_async(),
).await

Events & ServiceChecks

// Send a datadog event.
client.event("event title", "event text", AlertType::Warning, tags.as_ref());

// Send a datadog service check.
client.service_check(
    "myapp.service.check.name",
    ServiceCheckStatus::Critical,
    tags.as_ref(),
);

Pipeline

Multiple metrics can be sent to StatsD once using pipeline:

let mut pipe = client.pipeline():

// Increment a counter by 1
pipe.incr("some.counter");

// Decrement a counter by 1
pipe.decr("some.counter");

// Update a gauge
pipe.gauge("some.value", 12.0);

// Modify a counter by an arbitrary float.
pipe.count("some.counter", 511.0);

// Send a histogram value as a float.
pipe.histogram("some.histogram", 511.0);

// Set max UDP packet size if you wish, default is 512
pipe.set_max_udp_size(128);

// Send to StatsD
pipe.send(&client);

Pipelines are also helpful to make functions simpler to test, as you can pass a pipeline and be confident that no UDP packets will be sent.

License

Licenesed under the MIT License.