In order to avoid a hard dependency on GitHub during instance provisioning, we've chosen to use an S3-based copy of socorro-infra/puppet[0][1]. This is all well and good, but it means that every time we make a change in Puppet, we also need to update the archive. Currently this is an irritatingly manual process. It *must* be automated. I propose that we deal with this like we (will) deal with the Socorro repo: webhooks and DeadCI. [0] https://github.com/mozilla/socorro-infra/tree/master/puppet [1] https://s3-us-west-2.amazonaws.com/org.mozilla.crash-stats.packages-public/prov_cache/socorro-infra__puppet.tar.gz

The plan now is to only ship AMIs and not do this anymore.