Kepler

Kepler (Kubernetes-based Efficient Power Level Exporter) is an open-source observability project (sustainability observability) under the Cloud Native Computing Foundation that estimates and exports energy consumption metrics for Kubernetes workloads and nodes.

• Estimates per-workload and per-node power consumption in Kubernetes clusters (sustainability observability).
• Collects hardware performance counters, system metrics, and resource usage to infer energy usage (metrics collection).
• Exposes energy and power metrics through Prometheus-compatible endpoints (monitoring and telemetry).
• Integrates with Kubernetes via DaemonSet deployment and supports multi-node cluster monitoring (cloud-native infrastructure observability).
• Enables energy-aware analysis for capacity planning, cost management, and sustainability reporting (resource optimization and governance).

More About Kepler

Kepler (Kubernetes-based Efficient Power Level Exporter) addresses the problem of observing and quantifying the energy consumption of containerized workloads and cluster infrastructure in Kubernetes environments (sustainability observability). It focuses on providing power and energy metrics so platform teams, SREs, and sustainability stakeholders can analyze how applications and nodes consume energy across clusters.

The project runs as an agent, typically deployed as a DaemonSet on Kubernetes nodes, where it collects low-level telemetry such as hardware performance counters, Central Processing Unit (CPU) and memory utilization, and other system metrics (metrics collection). Kepler correlates this data to estimate power consumption per node and attribute it to individual pods or workloads, exposing these estimations as time-series metrics (workload-level energy accounting). This approach allows enterprises to gain visibility into power usage without requiring specialized power metering hardware on every node.

Kepler publishes its metrics through a Prometheus-compatible endpoint (monitoring and telemetry), which allows existing observability stacks that use Prometheus, Alertmanager, and related tools to ingest and visualize energy data. Metrics can be combined with application performance, cost, or Service Level Objective (SLO) dashboards to analyze energy efficiency in the context of throughput, latency, or infrastructure footprint (operations analytics). The project’s design aligns with common cloud-native patterns, including containerized deployment, Kubernetes manifests, and integration with service monitoring pipelines.

In enterprise environments, Kepler supports use cases such as comparing energy usage across namespaces, teams, or services, evaluating the energy effect of configuration changes, and tracking the energy characteristics of different hardware profiles or clusters (capacity planning and optimization). Organizations can export Kepler data into observability platforms or reporting pipelines to support sustainability programs, internal carbon accounting, and data center efficiency initiatives (sustainability reporting).

Kepler sits within the cloud-native ecosystem under the Cloud Native Computing Foundation and aligns with other CNCF projects that rely on Prometheus metrics and Kubernetes-native deployment patterns (cloud-native ecosystem integration). Its focus on power and energy metrics positions it in directories and taxonomies under sustainability observability, infrastructure monitoring, and energy-aware resource optimization. For technical stakeholders, Kepler provides a programmable and queryable source of energy metrics, enabling policy engines, autoscalers, and analytics tools to incorporate energy considerations alongside traditional performance and reliability metrics.