[Puppet Users] How does puppet fit into release cycle workflows

Hi there,

I am new to puppet and try to wrap my head around things. Puppet is
primarily for config management. You can however install packages with it.
I get all that. What I try to understand is how it fits into repository
lifecycle management. This is probably not the domain of puppet but I will
ask anyway because I am very interested how others do it. I am used to Red
Hat's Satellite way of life cycle management where one has repositories for
dev, qa, prod, etc. Package updates enter dev. Once OK-ed, they get cloned
to qa and then to prod. Hosts are assigned to these repo's. So once package
updates reach prod, we can update the prod servers with the tested
packages. The packages themselves are from Red Hat itself (security,
bugfix, and enhancement errata), or custom packages.

Now, I see some people saying they use Satellite (or spacewalk, it's
upstream product) for package management alone (although it can do config
file management) and puppet for configuration management. I would like to
use one tool for both package (update) installations and config changes.

So, how have you guys set up these workflows?

Tanks for any pointers!

Fred