Monitor dockerhub rate limits with grafana

Monitor dockerhub rate limits with grafana

AboutUsageImage ConfigurationLocal DemoHelm ChartTodoIssuesAcknowledgment


About

The exporter allows to retrieve the DockerHub rate limit counts as scrape target for Prometheus. The exporter obtains an auth token and then queries the Docker Hub registry with a HEAD request to parse RateLimit-Limit, RateLimit-Remaining and RateLimit-Reset into a Gauge metric. You can use your Docker Hub credentials to authenticate, otherwise an anonymous token is used.

Screenshot

Usage

Multi Arch docker images are available (arm/arm64/amd64) you can pull it from dockerhub and run in your environment.

1docker pull khaliq/drl-exporter:latest
2
3docker run -d -p 2121:2121  khaliq/drl-exporter:latest
4
5curl localhost:2121/metrics

Output

 1# HELP dockerhub_limit_max_requests_time Dockerhub rate limit maximum requests total time seconds
 2# TYPE dockerhub_limit_max_requests_time gauge
 3dockerhub_limit_max_requests_time 21600
 4# HELP dockerhub_limit_max_requests_total Dockerhub rate limit maximum requests in given time
 5# TYPE dockerhub_limit_max_requests_total gauge
 6dockerhub_limit_max_requests_total 100
 7# HELP dockerhub_limit_remaining_requests_time Dockerhub rate limit remaining requests time seconds
 8# TYPE dockerhub_limit_remaining_requests_time gauge
 9dockerhub_limit_remaining_requests_time 21600
10# HELP dockerhub_limit_remaining_requests_total Dockerhub rate limit remaining requests in given time
11# TYPE dockerhub_limit_remaining_requests_total gauge
12dockerhub_limit_remaining_requests_total 99

To build the image in your local environment
1git clone https://github.com/m47ik/drl-exporter.git
2cd drl-exporter
3make docker

Configuration Variables

Variables Default Value Discription
EXPORTER_PORT 2121 Server listening port
ENABLE_USER_AUTH false️ Must be set to true if providing username
DOCKERHUB_USER "" Dockerhub account
DOCKERHUB_PASSWORD "" Account password
DOCKERHUB_REPO_IMAGE ratelimitpreview/test custom repository/image

Local Demo

You can find the complete docker-compose file along with a dashboard under deploy folder to test it out.

1cd deploy/docker-compose
2docker-compose up -d

Web UI

Web URL
Grafana http://localhost:3000
Prometheus http://localhost:9090
Exporter http://localhost:8881

Helm Chart

  1. git clone https://github.com/m47ik/drl-exporter.git
  2. cd drl-exporter
  3. helm install <release name> deploy/chart --namespace=<desired namespace>

Installing chart with username and password

You can tweak the options for chart by setting values at run time or values.yaml file. If you intend to use the exporter with a username and password do remember to set the enableUserAuth=true as well.

1helm install my-release deploy/chart --set config.dockerhubUsername=<username>,
2config.dockerhubPassword=<password>,config.enableUserAuth=true  --namespace=<namespace>

Chart Configuration

Parameter Description Default
config.exporterPort Port the deployment exposes 2121
config.enableUserAuth Enable metrics for specific dockerhub account false
config.dockerhubUsername Dockerhub Username nil
config.dockerhubPassword Dockerhub Password nil
serviceMonitor.enabled If true, creates a ServiceMonitor instance false
serviceMonitor.additionalLabels Configure additional labels for the servicemonitor {}
serviceMonitor.namespace The namespace into which the servicemonitor is deployed. same as chart namespace
serviceMonitor.interval The interval with which prometheus will scrape 30s
serviceMonitor.scrapeTimeout The timeout for the scrape request 10s