@aws-lambda-powertools/metrics
Advanced tools
The metrics package for the Powertools for AWS Lambda (TypeScript) library
@aws-lambda-powertools/metrics is an npm package designed to simplify the creation and management of custom metrics in AWS Lambda functions. It provides utilities to easily capture and publish metrics to Amazon CloudWatch, enabling better observability and monitoring of serverless applications.
Creating Custom Metrics
This feature allows you to create custom metrics and publish them to CloudWatch. The code sample demonstrates how to add metrics for successful invocations and processing time, and then publish them.
const { Metrics, MetricUnits } = require('@aws-lambda-powertools/metrics');
const metrics = new Metrics();
exports.handler = async (event) => {
metrics.addMetric('SuccessfulInvocations', MetricUnits.Count, 1);
metrics.addMetric('ProcessingTime', MetricUnits.Milliseconds, 200);
await metrics.publishStoredMetrics();
return { statusCode: 200, body: 'Metrics published' };
};Namespace and Dimensions
This feature allows you to set a namespace and dimensions for your metrics, which helps in organizing and filtering them in CloudWatch. The code sample shows how to set a namespace and dimensions for a payment service.
const { Metrics, MetricUnits } = require('@aws-lambda-powertools/metrics');
const metrics = new Metrics({ namespace: 'MyApp', dimensions: { Service: 'Payment' } });
exports.handler = async (event) => {
metrics.addMetric('PaymentSuccess', MetricUnits.Count, 1);
await metrics.publishStoredMetrics();
return { statusCode: 200, body: 'Payment metrics published' };
};Automatic Cold Start Metric
This feature automatically captures a cold start metric, which is useful for understanding the performance impact of cold starts in your Lambda functions. The code sample demonstrates how to enable this feature.
const { Metrics, MetricUnits } = require('@aws-lambda-powertools/metrics');
const metrics = new Metrics({ captureColdStartMetric: true });
exports.handler = async (event) => {
metrics.addMetric('HandlerInvoked', MetricUnits.Count, 1);
await metrics.publishStoredMetrics();
return { statusCode: 200, body: 'Cold start metric captured' };
};The aws-sdk package is the official AWS SDK for JavaScript, which includes support for CloudWatch metrics. However, it requires more boilerplate code to achieve the same functionality as @aws-lambda-powertools/metrics. It is more general-purpose and not specifically tailored for Lambda functions.
The cloudwatch-metrics package is a lightweight library for publishing metrics to CloudWatch. It provides a simpler interface compared to the aws-sdk but lacks some of the advanced features and integrations offered by @aws-lambda-powertools/metrics, such as automatic cold start metrics.
Powertools for AWS Lambda (TypeScript) is a developer toolkit to implement Serverless best practices and increase developer velocity.
You can use the library in both TypeScript and JavaScript code bases.
The library provides a utility function to emit metrics to CloudWatch using Embedded Metric Format (EMF).
To get started, install the library by running:
npm i @aws-lambda-powertools/metrics
After initializing the Metrics class, you can add metrics using the https://docs.powertools.aws.dev/lambda/typescript/latest/core/metrics/#creating-metrics method. The metrics are stored in a buffer and are flushed when calling https://docs.powertools.aws.dev/lambda/typescript/latest/core/metrics/#flushing-metrics.
Each metric can also have dimensions and metadata added to it.
import { Metrics, MetricUnit } from '@aws-lambda-powertools/metrics';
const metrics = new Metrics({
namespace: 'serverlessAirline',
serviceName: 'orders',
defaultDimensions: { environment: process.env.ENVIRONMENT },
});
export const handler = async (event: { requestId: string }) => {
metrics.addMetadata('request_id', event.requestId);
metrics.addMetric('successfulBooking', MetricUnit.Count, 1);
metrics.publishStoredMetrics();
};
As you finish adding all your metrics, you need to serialize and "flush them" by calling publishStoredMetrics(), which will emit the metrics to stdout in the Embedded Metric Format (EMF). The metrics are then picked up by the Lambda runtime and sent to CloudWatch.
The publishStoredMetrics() method is synchronous and will block the event loop until the metrics are flushed. If you want Metrics to flush automatically at the end of your Lambda function, you can use the @logMetrics() decorator or the logMetrics() middleware.
With Metrics, you can capture cold start as a metric by calling the captureColdStartMetric() method. This method will add a metric with the name ColdStart and the value 1 to the metrics buffer.
This metric is flushed automatically as soon as the method is called, to ensure that the cold start is captured regardless of whether the metrics are flushed manually or automatically.
import { Metrics, MetricUnit } from '@aws-lambda-powertools/metrics';
const metrics = new Metrics({
namespace: 'serverlessAirline',
serviceName: 'orders',
});
export const handler = async (event: { requestId: string }) => {
metrics.captureColdStartMetric();
};
Note that we don't emit a ColdStart metric with value 0 when the function is warm, as this would result in a high volume of metrics being emitted to CloudWatch, so you'll need to rely on the absence of the ColdStart metric to determine if the function is warm.
If you are using TypeScript and are comfortable with writing classes, you can use the @logMetrics() decorator to automatically flush metrics at the end of your Lambda function as well as configure additional options such as throwing an error if no metrics are added, capturing cold start as a metric, and more.
import type { Context } from 'aws-lambda';
import type { LambdaInterface } from '@aws-lambda-powertools/commons/types';
import { Metrics, MetricUnit } from '@aws-lambda-powertools/metrics';
const metrics = new Metrics({
namespace: 'serverlessAirline',
serviceName: 'orders',
});
class Lambda implements LambdaInterface {
@metrics.logMetrics({ captureColdStartMetric: true, throwOnEmptyMetrics: true })
public async handler(event: { requestId: string }, _: Context) {
metrics.addMetadata('request_id', event.requestId);
metrics.addMetric('successfulBooking', MetricUnit.Count, 1);
}
}
const handlerClass = new Lambda();
export const handler = handlerClass.handler.bind(handlerClass);
Decorators are a Stage 3 proposal for JavaScript and are not yet part of the ECMAScript standard. The current implmementation in this library is based on the legacy TypeScript decorator syntax enabled by the experimentalDecorators flag set to true in the tsconfig.json file.
If instead you are using Middy.js and prefer to use middleware, you can use the @logMetrics() middleware to do the same as the class method decorator.
The @logMetrics() middleware can be used with Middy.js to automatically flush metrics at the end of your Lambda function as well as configure additional options such as throwing an error if no metrics are added, capturing cold start as a metric, and set default dimensions.
import { Metrics, MetricUnit } from '@aws-lambda-powertools/metrics';
import { logMetrics } from '@aws-lambda-powertools/metrics/middleware';
import middy from '@middy/core';
const metrics = new Metrics({
namespace: 'serverlessAirline',
serviceName: 'orders',
});
export const handler = middy(async (event) => {
metrics.addMetadata('request_id', event.requestId);
metrics.addMetric('successfulBooking', MetricUnit.Count, 1);
}).use(logMetrics(metrics, {
captureColdStartMetric: true,
throwOnEmptyMetrics: true,
}));
The logMetrics() middleware is compatible with @middy/core@3.x and above.
If you are interested in contributing to this project, please refer to our Contributing Guidelines.
The roadmap of Powertools for AWS Lambda (TypeScript) is driven by customers’ demand.
Help us prioritize upcoming functionalities or utilities by upvoting existing RFCs and feature requests, or creating new ones, in this GitHub repository.
#typescript - Invite linkKnowing which companies are using this library is important to help prioritize the project internally. If your company is using Powertools for AWS Lambda (TypeScript), you can request to have your name and logo added to the README file by raising a Support Powertools for AWS Lambda (TypeScript) (become a reference) issue.
The following companies, among others, use Powertools:
Share what you did with Powertools for AWS Lambda (TypeScript) 💞💞. Blog post, workshops, presentation, sample apps and others. Check out what the community has already shared about Powertools for AWS Lambda (TypeScript) here.
This helps us understand who uses Powertools for AWS Lambda (TypeScript) in a non-intrusive way, and helps us gain future investments for other Powertools for AWS Lambda languages. When using Layers, you can add Powertools as a dev dependency to not impact the development process.
This library is licensed under the MIT-0 License. See the LICENSE file.
2.12.0 (2024-12-17)
POWERTOOLS_METRICS_DISABLED (#3351) (7e8578e)FAQs
The metrics package for the Powertools for AWS Lambda (TypeScript) library
The npm package @aws-lambda-powertools/metrics receives a total of 123,399 weekly downloads. As such, @aws-lambda-powertools/metrics popularity was classified as popular.
We found that @aws-lambda-powertools/metrics demonstrated a healthy version release cadence and project activity because the last version was released less than a year ago. It has 3 open source maintainers collaborating on the project.
Did you know?
Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.
Research
Security News
Socket researchers uncover the risks of a malicious Python package targeting Discord developers.
Security News
The UK is proposing a bold ban on ransomware payments by public entities to disrupt cybercrime, protect critical services, and lead global cybersecurity efforts.
Security News
Snyk's use of malicious npm packages for research raises ethical concerns, highlighting risks in public deployment, data exfiltration, and unauthorized testing.