In this chapter, the word PuppetDB has been used zillions of times as an obsessive mantra. While we can use Puppet without it, as we have done for years, it's important to realize that PuppetDB is going to be present in every relevant Puppet infrastructure, and we can bet more and more applications and tools will emerge around it.

The fact that it's a robust piece of software engineered in a brilliant way makes us feel comfortable about the idea that Puppet Labs has decided to use it as central point of consolidation and gathering for all the data generated by Puppet.

We have seen how to configure PuppetDB and its integration with the Puppet Master, and how to interpret its performance dashboard. We have explored the principles of PuppetDB CQRS API, with REST-like endpoints for queries and commands for writing and, in some detail, the list of available endpoints, with various sample queries.

Finally, we have seen how most of the information gathered by PuppetDB can be queried from our manifests using the puppetdbquery module, and how this can dramatically influence how we manage interactions among different nodes.

Now that we grasp firmly the principles of Puppet, Hiera, and PuppetDB, we can explore how they can be glued together. In the next chapter, we will see how to deploy different architectures with them.