The interface is not stable
Provide a backend for the metrics
facade crate,
to emit metrics in CloudWatch Embedded Metrics Format
```rust let metrics = metricscloudwatchembedded::Builder::new() .cloudwatch_namespace("MyApplication") .init() .unwrap();
metrics::increment_counter!("requests", "Method" => "Default");
metrics .set_property("RequestId", "ABC123") .flush(std::io::stdout()); ```
The Lambda Runtime intergration feature handles flushing metrics
after each invoke via either run()
alternatives or MetricService
which inplements the
tower::Service
trait.
It also provides optional helpers for:
* emiting a metric on cold starts
* wrapping cold starts in a tracing
span
* decorating metric documents with request id and/or x-ray trace id
In your Cargo.toml add:
toml
metrics = "0.21"
metrics_cloudwatch_embedded = { version = "0.4.0", features = ["lambda"] }
tracing-subscriber = { version = "0.3", default-features = false, features = ["fmt", "env-filter", "json"] }
main.rs: ```rust use lambdaruntime::{Error, LambdaEvent}; use metricscloudwatch_embedded::lambda::handler::run; use serde::{Deserialize, Serialize}; use tracing::{info, span, Level};
struct Request {}
struct Response { req_id: String, }
async fn functionhandler(event: LambdaEvent
info!("Hello from function_handler");
metrics::increment_counter!("requests", "Method" => "Default");
Ok(resp)
}
async fn main() -> Result<(), Error> { tracingsubscriber::fmt() .json() .withenvfilter(tracingsubscriber::filter::EnvFilter::fromdefaultenv()) .withtarget(false) .withcurrentspan(false) .withouttime() .init();
let metrics = metrics_cloudwatch_embedded::Builder::new()
.cloudwatch_namespace("MetricsTest")
.with_dimension("Function", std::env::var("AWS_LAMBDA_FUNCTION_NAME").unwrap())
.lambda_cold_start_span(span!(Level::INFO, "cold start").entered())
.lambda_cold_start_metric("ColdStart")
.with_lambda_request_id("RequestId")
.init()
.unwrap();
info!("Hello from main");
run(metrics, function_handler).await
}
CloudWatch log after a single invoke (cold start):
plaintext
INITSTART Runtime Version: provided:al2.v19 Runtime Version ARN: arn:aws:lambda:us-west-2::runtime:d1007133cb0d993d9a42f9fc10442cede0efec65d732c7943b51ebb979b8f3f8
{"level":"INFO","fields":{"message":"Hello from main"},"spans":[{"name":"cold start"}]}
START RequestId: fce53486-160d-41e8-b8c3-8ef0fd0f4051 Version: $LATEST
{"aws":{"Timestamp":1688294472338,"CloudWatchMetrics":[{"Namespace":"MetricsTest","Dimensions":[["Function"]],"Metrics":[{"Name":"ColdStart","Unit":"Count"}]}]},"Function":"MetricsTest","RequestId":"fce53486-160d-41e8-b8c3-8ef0fd0f4051","ColdStart":1}
{"level":"INFO","fields":{"message":"Hello from functionhandler"},"spans":[{"name":"cold start"},{"requestId":"fce53486-160d-41e8-b8c3-8ef0fd0f4051","xrayTraceId":"Root=1-64a15448-4aa914a00d66aa066325d7e3;Parent=60a7d0c22fb2f001;Sampled=0;Lineage=16f3a795:0","name":"Lambda runtime invoke"}]}
{"aws":{"Timestamp":1688294472338,"CloudWatchMetrics":[{"Namespace":"MetricsTest","Dimensions":[["Function","Method"]],"Metrics":[{"Name":"requests"}]}]},"Function":"MetricsTest","Method":"Default","RequestId":"fce53486-160d-41e8-b8c3-8ef0fd0f4051","requests":1}
END RequestId: fce53486-160d-41e8-b8c3-8ef0fd0f4051
REPORT RequestId: fce53486-160d-41e8-b8c3-8ef0fd0f4051 Duration: 1.22 ms Billed Duration: 11 ms Memory Size: 128 MB Max Memory Used: 13 MB Init Duration: 8.99 ms
```
collector::Collector::flush
, overflow will report an error via the tracing
crateBuilder::with_dimension(...)
may not overlap with metric labels
metrics::Unit
are supported
https://docs.aws.amazon.com/AmazonCloudWatch/latest/APIReference/API_MetricDatum.htmlmetrics::Key
will fail with an error via the tracing
cratetracing
crateThe AWS Lambda Rust Runtime requires a minimum of Rust 1.62, and is not guaranteed to build on compiler versions earlier than that.
This may change when async traits are released to stable depending on the ripple effects to the ecosystem.
This project is licensed under the Apache-2.0 License. Apache-2.0 was chosen to match the Lambda Runtime
Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be licensed as above, without any additional terms or conditions.