minio/docs/metrics/prometheus

How to monitor MinIO server with Prometheus

Prometheus is a cloud-native monitoring platform.

Prometheus offers a multi-dimensional data model with time series data identified by metric name and key/value pairs. The data collection happens via a pull model over HTTP/HTTPS.

MinIO exports Prometheus compatible data by default as an authorized endpoint at /minio/prometheus/metrics/cluster.

Users looking to monitor their MinIO instances can point Prometheus configuration to scrape data from this endpoint.

This document explains how to setup Prometheus and configure it to scrape data from MinIO servers.

Table of Contents

• Prerequisites
  • 1. Download Prometheus
  • 2. Configure authentication type for Prometheus metrics
  • 3. Configuring Prometheus
    • 3.1 Authenticated Prometheus config
    • 3.2 Public Prometheus config
  • 4. Update scrape_configs section in prometheus.yml
  • 5. Start Prometheus
  • 6. Configure Grafana
• List of metrics exposed by MinIO

Prerequisites

To get started with MinIO, refer MinIO QuickStart Document. Follow below steps to get started with MinIO monitoring using Prometheus.

1. Download Prometheus

Download the latest release of Prometheus for your platform, then extract it

tar xvfz prometheus-*.tar.gz
cd prometheus-*

Prometheus server is a single binary called prometheus (or prometheus.exe on Microsoft Windows). Run the binary and pass --help flag to see available options

./prometheus --help
usage: prometheus [<flags>]

The Prometheus monitoring server

. . .

Refer Prometheus documentation for more details.

2. Configure authentication type for Prometheus metrics

MinIO supports two authentication modes for Prometheus either jwt or public, by default MinIO runs in jwt mode. To allow public access without authentication for prometheus metrics set environment as follows.

export MINIO_PROMETHEUS_AUTH_TYPE="public"
minio server ~/test

3. Configuring Prometheus

3.1 Authenticated Prometheus config

If MinIO is configured to expose metrics without authentication, you don't need to use mc to generate prometheus config. You can skip reading further and move to 3.2 section.

The Prometheus endpoint in MinIO requires authentication by default. Prometheus supports a bearer token approach to authenticate prometheus scrape requests, override the default Prometheus config with the one generated using mc. To generate a Prometheus config for an alias, use mc as follows mc admin prometheus generate <alias>.

The command will generate the scrape_configs section of the prometheus.yml as follows:

scrape_configs:
- job_name: minio-job
  bearer_token: <secret>
  metrics_path: /minio/v2/metrics/cluster
  scheme: http
  static_configs:
  - targets: ['localhost:9000']

3.2 Public Prometheus config

If Prometheus endpoint authentication type is set to public. Following prometheus config is sufficient to start scraping metrics data from MinIO. This can be collected from any server once per collection.

Cluster

scrape_configs:
- job_name: minio-job
  metrics_path: /minio/v2/metrics/cluster
  scheme: http
  static_configs:
  - targets: ['localhost:9000']

Node

Optionally you can also collect per node metrics. This needs to be done on a per server instance.

scrape_configs:
- job_name: minio-job
  metrics_path: /minio/v2/metrics/node
  scheme: http
  static_configs:
  - targets: ['localhost:9000']

4. Update scrape_configs section in prometheus.yml

To authorize every scrape request, copy and paste the generated scrape_configs section in the prometheus.yml and restart the Prometheus service.

5. Start Prometheus

Start (or) Restart Prometheus service by running

./prometheus --config.file=prometheus.yml

Here prometheus.yml is the name of configuration file. You can now see MinIO metrics in Prometheus dashboard. By default Prometheus dashboard is accessible at http://localhost:9090.

6. Configure Grafana

After Prometheus is configured, you can use Grafana to visualize MinIO metrics. Refer the document here to setup Grafana with MinIO prometheus metrics.

List of metrics exposed by MinIO

MinIO server exposes the following metrics on /minio/prometheus/metrics/cluster endpoint. All of these can be accessed via Prometheus dashboard. A sample list of exposed metrics along with their definition is available in the demo server at curl https://play.min.io:9000/minio/prometheus/metrics/cluster

List of metrics reported

The list of metrics reported can be here