# Prometheus nvidia-smi Exporter

Prometheus nvidia-smi Exporter is a self-hosted monitoring & logging app. A lightweight Unraid plugin that exposes GPU statistics from nvidia-smi as Prometheus metrics, making the data available to any monitoring stack that can scrape a standard /metrics endpoint. It is open source and self-hosted, installable on Unraid.

Canonical page: https://ourlocalhost.com/app/4113/prometheus-nvidia-smi-exporter

## Description

A lightweight Unraid plugin that exposes GPU statistics from nvidia-smi as Prometheus metrics, making the data available to any monitoring stack that can scrape a standard /metrics endpoint.

## Platform availability

Available on 1 of 10 platforms: Unraid.

## Health score

25/100 (Low), based on 4 of 5 factors.
- Maintained: 0 (last commit 732 days ago)
- Popular: 26 (18 GitHub stars)
- Easy to install: 25 (1 install platform, install notes)
- Security check: 75 (no serious findings)

## Links

- Unraid: https://ca.unraid.net/apps/prometheus-nvidia-smi-exporter-1nhczuk150lf34?q=Prometheus%20nvidia-smi%20Exporter
- GitHub: https://github.com/e7d/docker-prometheus-nvidiasmi

## Common questions

### What Docker settings are required for the exporter to access the GPU?

Run the container in privileged mode with the nvidia runtime, for example --privileged --runtime nvidia.

### Where should the exporter be added in Prometheus?

Add the Unraid host IP with port 9202 as the scrape target. Metrics are served at /metrics.

### Does the exporter support multiple GPUs?

Yes, the exporter supports multiple GPUs.

### Are GPU metrics limited on a DGX Spark GB10?

GPU memory and process metrics may be limited on a DGX Spark GB10 because the GPU has no discrete VRAM.

## Alternatives

- Amazon CloudWatch: Amazon CloudWatch is a monitoring service for AWS cloud resources and the applications running on AWS.
- Better Stack: AI-driven monitoring platform providing incident management, real-time logs, uptime, traces, error tracking, infrastructure analytics, and…
- Datadog: Datadog Agent is an on-host monitoring agent that collects metrics, logs, and traces from hosts, containers, and Kubernetes clusters and se…
- DeepFlow: Unify your cloud-native monitoring with zero-code eBPF, advanced profiling, and 10x storage efficiency.
- Grafana Loki: Grafana Loki is an open-source log aggregation system.
- Graylog: Graylog is a centralized log management platform built on open standards for capturing, storing, and analyzing terabytes of machine data in…
- InfluxDB: InfluxDB is an open source time series database with no external dependencies.
- Kibana: Kibana is the data visualization layer for querying, analyzing, and managing data stored in Elasticsearch.
- New Relic: Pinpoint and solve performance issues in your Ruby, Java, .NET, PHP and Python apps with real user, application and server monitoring.
- Prometheus: Prometheus is a systems and service monitoring system.
- Splunk: Splunk is a self-hosted platform for collecting, indexing, searching, and analyzing log and security data, helping homelabbers and operatio…
- VictoriaMetrics: VictoriaMetrics is a fast, cost-effective, and scalable open-source time-series database that stores long-term monitoring metrics, accepts…

---
Source: Our Local Host (https://ourlocalhost.com/app/4113/prometheus-nvidia-smi-exporter), the self-hosted and open-source app directory.
