aws-embedded-metrics

What it is and what it does

aws-embedded-metrics is a Python library for publishing custom CloudWatch metrics alongside structured log events. It wraps metric emission in the Embedded Metric Format (EMF), which CloudWatch automatically extracts so you can visualize, alarm on, and aggregate metrics in real time without writing custom batching code or making blocking network requests. The library automatically injects environment metadata (Lambda version, EC2 instance IDs, etc.) into logs and supports both standard and high-resolution metrics.

You decorate your function with @metric_scope, then call methods like put_metric(), put_dimensions(), and set_property() on the metrics object to record values and context. The library handles flushing metrics to CloudWatch and formatting them according to EMF. It works across Lambda, EC2, ECS, EKS, and on-premises environments (the latter via CloudWatch Agent), and integrates with CloudWatch Logs Insights for querying high-cardinality context that wouldn't be suitable as metric dimensions.

Use it for:

• Emit custom metrics from Lambda functions without custom batching or blocking network calls
• Link metrics to high-cardinality context (request IDs, device IDs) queryable via CloudWatch Logs Insights
• Monitor aggregated values across EC2, ECS, EKS, or on-premises compute while preserving detailed event logs
• Track both standard (1-minute) and high-resolution (sub-minute) metrics from application code
• Automatically capture environment metadata (Lambda version, instance IDs) alongside custom metrics

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Generates CloudWatch metrics embedded within structured log events, allowing you to visualize and alarm on custom metrics while preserving detailed event context for querying.

Yes. The package is production-stable, actively maintained, has low install friction, permissive licensing, and no known vulnerabilities. It solves a real AWS monitoring problem—emitting custom metrics without blocking calls or external dependencies—and is widely used (top 5000 PyPI packages). Install it if you need to publish CloudWatch metrics from Lambda or other AWS compute environments.

Install

aws-embedded-metrics on PyPI

pip install aws-embedded-metrics

uv add aws-embedded-metrics

poetry add aws-embedded-metrics

Installing aws-embedded-metrics

Before you install

Low install friction with a single runtime dependency (aiohttp). Actively maintained with recent commits and production-stable status. Supports current Python versions (3.6+).

License in practice

Licensed under Apache Software License (permissive), allowing commercial and private use with minimal restrictions.

Quickstart

pip install aws-embedded-metrics

from aws_embedded_metrics import metric_scope
from aws_embedded_metrics.storage_resolution import StorageResolution

@metric_scope
def my_handler(metrics):
    metrics.put_dimensions({"Foo": "Bar"})
    metrics.put_metric("ProcessingLatency", 100, "Milliseconds", StorageResolution.STANDARD)
    metrics.set_property("RequestId", "422b1569-16f6-4a03")
    return {"message": "Hello!"}

Verify before relying

• Whether aiohttp is required for all use cases or only for async operations
• Performance characteristics with high-throughput generators and the v4.0 flush behavior change

Package facts

Evidence: aws_embedded_metrics-3.5.0-py3-none-any.whl

Keywords: aws, logs, metrics, emf

Tags

More Monitoring packages