Centralized server administration and monitoring

Part of our ongoing goal with Ubuntu Server is to simplify the system administrator's life, and providing methods to centralize administration and monitoring in various ways are a good way of doing that.

We should choose a system to provide monitoring (such as ganglia) and make sure it can be installed and configured with minimal hassle. Currently such systems are relatively non-trivial to configure and deploy.

As for centralized administration, we should review the etc-in-svn spec (misleading name; VCS is actually bzr) to make sure it agrees with how we want to handle configuration versioning. We should also review nwu, the network-wide updates tool being written, for design and security issues, and hopefully endorse it for future inclusion in Ubuntu Server.

Thoughts about general server farm partitioning and administration should be written down to form an informational roadmap.