opentelemetry-tide

为您的tide应用程序添加OpenTelemetry跟踪和指标支持。成为新可观察性运动的成员！

注意

• 它只实现了中间件层上非常基本的请求跟踪。如果您需要执行代码的跨度，您需要自己添加。
• 它提供了基于RED方法的基本Prometheus指标。
• 该项目受到了https://github.com/OutThereLabs/actix-web-opentelemetry的启发。
• 您可能不想在生产环境中使用它。🤷

如何使用

# Run jaeger in background
docker run -d \
  -p6831:6831/udp -p6832:6832/udp -p16686:16686 -p14268:14268 \
  jaegertracing/all-in-one:latest

# Run server example with tracing middleware
cargo run --example server

# Make a request or two ...
curl https://127.0.0.1:3000/

# Open browser and view the traces
firefox https://127.0.0.1:16686/

# Check the prometheus metrics endpoint
curl https://127.0.0.1:3000/metrics

示例

Cargo.toml

# ...

[dependencies]
async-std = { version = "1.10", features = ["attributes"] }
opentelemetry = { version = "0.17.0", features = ["rt-async-std"] }
opentelemetry-jaeger = { version = "0.16.0", features = ["rt-async-std"] }
opentelemetry-tide = "0.12"
tide = "0.16"

server.rs

use opentelemetry::{global, KeyValue, runtime};
use opentelemetry_semantic_conventions::resource;
use opentelemetry_tide::TideExt; // import trait

const VERSION: &'static str = env!("CARGO_PKG_VERSION");

#[async_std::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    tide::log::with_level(tide::log::LevelFilter::Warn);

    let tags = [resource::SERVICE_VERSION.string(VERSION)];

    let _tracer = opentelemetry_jaeger::new_pipeline()
        .with_tags(tags.iter().map(ToOwned::to_owned))
        .install_batch(runtime::AsyncStd)
        .expect("pipeline install failure");
    let tracer = global::tracer("example-server");

    let metrics_kvs = vec![KeyValue::new("K", "V")];

    let mut app = tide::new();

    // use the trait
    app.with_middlewares(tracer, Some(metrics_kvs));

    app.at("/").get(|_| async move {
        Ok("Hello, OpenTelemetry!")
    });

    app.listen("0.0.0.0:3000").await?;

    global::shutdown_tracer_provider();

    Ok(())
}

Cargo功能

安全性

此包使用 #![forbid(unsafe_code)] 确保一切都在100%安全的Rust中实现。

许可证