Stackalytics was created in 2013 by Mirantis in order to provide a reference source of contribution metrics for the OpenStack community. In that year all companies were trying to position themselves as the #1 OpenStack company, so people were very interested by such information. It's now an infra project which replaced activity.openstack.org, and the infra team had plans to host it directly, at stackalytics.openstack.org.
However, as OpenStack grew, adding up and comparing contributions between vastly-different projects (code, doc, config) using simple metrics (commits, reviews, LOC) looked more and more like comparing apples to oranges. Without deeper analysis, the base results are very misleading. Moreover, its status of reference site incentivized people to optimize their contribution in terms of Stackalytics reports, rather than where the project really needed help.
At the same time, Stackalytics is now mostly unmaintained, with standing bugs which cause inaccurate results for some projects, duplication of some data, and load issues on review.openstack.org. Nobody so far picked up the spec to move it under OpenStack infrastructure.
Is it time for Stackalytics to die ? Is such a reference site, in its current state, more harmful than helpful ? Should raw unfiltered metrics be replaced by thoughtful analysis digging deeper into the data and trends ? The Linux kernel is happy with after-the-fact analysis provided by Jon Corbet at LWN. Would a community report, produced by the OpenStack Foundation every release (and every year) be sufficient to give a state of the contribution to OpenStack projects ?
Moderator: Thierry Carrez
