A Prometheus exporter that fetches metrics from Amazon CloudWatch and exposes them for monitoring.
CloudWatch Exporter is a Prometheus exporter that retrieves metrics from Amazon CloudWatch and exposes them in a format that Prometheus can scrape. It solves the problem of integrating AWS cloud metrics into a Prometheus-based monitoring stack, allowing teams to monitor AWS resources alongside other infrastructure components. The tool supports flexible configuration for selecting specific metrics, dimensions, and tags, and can be deployed as a standalone Java application or Docker container.
DevOps engineers, SREs, and cloud administrators who use Prometheus for monitoring and need to include AWS CloudWatch metrics in their observability pipeline. It is particularly useful for organizations running hybrid or multi-cloud environments where unified monitoring is essential.
Developers choose CloudWatch Exporter because it is an official Prometheus community project that provides a stable and well-maintained bridge to AWS metrics. Its configurable filtering and tag-based selection reduce unnecessary API calls, helping control costs, while support for Docker and live reloading simplifies deployment and management.
Metrics exporter for Amazon AWS CloudWatch
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
As a Prometheus community project, it offers stable and reliable bridging between CloudWatch and Prometheus, ensuring long-term support and compatibility.
Supports YAML-based configs with dimension filtering, tag-based selection (aws_tag_select), and per-metric overrides, allowing precise control over metric collection.
Includes features like dimension caching and experimental GetMetricData API support to optimize API calls and reduce CloudWatch costs, with built-in metrics to track requests.
Provides pre-built Docker images from prom/cloudwatch-exporter, simplifying deployment in containerized environments like Kubernetes.
Default 600-second delay to wait for CloudWatch data convergence means metrics are not real-time, requiring offset queries in PromQL for accurate viewing.
Requires in-depth knowledge of AWS metric dimensions and YAML syntax for configuration, with no automatic discovery, leading to steep initial setup effort.
Depends on Java 11, adding memory and performance overhead compared to lighter exporters, which can be a burden in resource-constrained setups.
Key performance improvements like GetMetricData are marked experimental and opt-in, so users must trade stability for better scalability.