Skip to content


GRid ACcounting Collector

(pronounced "grok")

GRACC is a collection of components for implementing resource usage accounting.


GRACC is meant as a replacement of the Gratia accounting system; the engineering focus is on smaller, independent components rather than a monolithic collector architecture. The hope is that, by breaking the functionality into a series of smaller components, future architectural changes (such as migration to a new database) can be done without rewriting the entire infrastructure. For example, forwarding information to a separate accounting database becomes much simpler in this infrastructure.

Repositories of interest:

  • GRACC Collector. An agent which runs on an existing Gratia collector that forwards raw usage records to GRACC.
  • GRACC Monitoring Emails. Simple daily emails overview GRACC activity.
  • GRACC Request Agent. Agents that listen for replay requests from GRACC.
  • GRACC Summary Agent. Agents that request summary records from the Request Agent and forward the records back to the Logstash agent for storage in ElasticSearch
  • GRACC Archiver. Agents that listen on the RabbitMQ exchange for raw records and store them to disk in a gzip file.
  • GRACC Backup Scripts. A service that runs periodically on the same host as the Archiver which backs up the completed gzip files to external Tape. See Backup Docs for more details.
  • GRACC Reports. A collection of reports that run periodically that query GRACC and present aggregated data to stakeholders.