Skip to content


Repository files navigation

Ceph RADOSGW Usage Exporter

Prometheus exporter that scrapes Ceph RADOSGW usage information (operations and buckets). This information is gathered from a RADOSGW using the Admin Operations API.

This exporter was based off from both ( and the more elaborate Jenkins exporter here (


  • Working Ceph Cluster with Object Gateways setup.
  • Ceph RADOSGWs must beconfigured to gather usage information as this is not on by default. The miniumum is to enable it via ceph.conf as below. There are however other options that are available and should be considered here. If you don't configure thresholds, intervals, and shards you may end up having too large objects in the usage namespace of the log pool. The values below are just examples. Check the documentation which ones would be the best ones for your setup.
rgw enable usage log = true
rgw usage log flush threshold = 1024
rgw usage log tick interval = 30
rgw usage max shards = 32
rgw usage max user shards = 8

  • Configure admin entry point (default is 'admin'):
rgw admin entry = "admin"
  • Enable admin API (default is enabled):
rgw enable apis = "s3, admin"
  • This exporter requires a user that has the following capability, see the Admin Guide here for more details.
    "caps": [
            "type": "buckets",
            "perm": "read"
            "type": "metadata",
            "perm": "read"
            "type": "usage",
            "perm": "read"
            "type": "users",
            "perm": "read"

Note: If using a loadbalancer in front of your RADOSGWs, please make sure your timeouts are set appropriately as clusters with a large number of buckets, or large number of users+buckets could cause the usage query to exceed the loadbalancer timeout.

For haproxy the timeout in question is timeout server

Local Installation

git clone
cd radosgw_usage_exporter
pip install requirements.txt


Arg Env Description Default
-H --host RADOSGW_SERVER Server URL for the RADOSGW api (example: http://radosgw:80
-e --admin-entry ADMIN_ENTRY The entry point for an admin request URL admin
-a --access-key ACCESS_KEY S3 access key NA
-s --secret-key SECRET_KEY S3 secret key NA
-k --insecure Allow insecure server connections when using SSL false
-p --port VIRTUAL_PORT Port to listen 9242
-S --store STORE Store name added to metrics us-east-1
-t --timeout TIMEOUT Timeout when getting metrics 60
-l --log-level LOG_LEVEL Provide logging level: DEBUG, INFO, WARNING, ERROR or CRITICAL INFO


./check_ceph_rgw_api -H -a JXUABTZZYHAFLCMF9VYV -s jjP8RDD0R156atS6ACSy2vNdJLdEPM0TJQ5jD1pw


Docker build (

docker run -d -p 9242 \

Arguments can also be specified by environment variables as well.

docker run -d -p 9242:9242 \
-e "RADOSGW_SERVER=<host>" \
-e "VIRTUAL_PORT=9242" \
-e "ACCESS_KEY=<access_key>" \
-e "SECRET_KEY=<secret_key>" \

Resulting metrics can be then retrieved via your Prometheus server via the http://<exporter host>:9242/metrics endpoint.


You can find an example of deployment using Rook operator in a K8s environment in examples/k8s directory.