Hoping to have some sort of download statistics on server side in terms of the builds and such so that we can utilize the builds better and see which builds aren't being used at all. ie Data driven methodology in regards to supporting well used builds and disabling indexes that aren't needed at all. Req: 1) hits via indexes/name space 2) hits via artifacts
We have GA for tools site.. But naturally we can't do GA serverside. Well, maybe but it doesn't make sense for automated downloads without a user-agent... We might already have some stats though. Are you think per artifact download stats? Or total number of downloads of any artifact with given name across all builds?
Basically, we don't think taskcluster should count artifact requests. But if we really wanted this feature, we could probably build something that could scan the access logs for a set of specific file names and aggregates statistics on that... However, it would be a rather significant investment to do this. I'm not inherently convinced it's something we should undertake. Things served to end-users should generally be served directly from a CDN instead of using taskcluster artifacts.. So this would only be relevant for small scale experiments, and perhaps such experiments should gather statistics by other means. Ie. a simple counter service that increments a counter and redirect to an artifact might be a better option. Certainly easier to build. -- This bug is closed as part of the Great Bugzilla Cleaning Sprint of 2016 by bstack and jonasfj, please reopen if you disagree.
Status: NEW → RESOLVED
Last Resolved: 2 years ago
Resolution: --- → WONTFIX
You need to log in before you can comment on or make changes to this bug.