Configuration Remediation

There was previously some requests for RHQ to support "configuration remediation". In short, it would be nice to "freeze" a resource configuration such that if that configuration ever changes for whatever reason, we want that configuration to immediately revert back to its "frozen" configuration state.

I spent a couple of hours coding up a prototype to show this feature in RHQ. See the flash demo here.

There are a few things to say about this.

First, specifically about configuration remediation - this demonstrates how easily it was for RHQ to support such a feature.

Second, and more generally, this effort illustrates the benefits of the RHQ architecture and the concept of "management at the edges". The code was added to the core configuration subsystem - it did not involve any one specific type of managed resource. Because of that, the couple hours that I spent adding this feature provided a wide array of functionality - I now have configuration remediation support for any managed resource that supports configuration. The storage of the configuration on the managed resources is handled at the edges by each manage resource's plugin. But the remediation logic is in the core server, independent of any managed resource that is out in the environment. So, not only was I able to quickly add a new configuration feature to RHQ, but that feature was immediately available across all types of resources without any additional effort.

What does this mean? It means, for example, that I can now freeze configuration of a JBossAS data source just as easily as I can freeze the configuration of my OpenSSH daemon (or any other managed resource that I can configure). On top of that, my RHQ users won't have to learn many different user interfaces or different mechanisms depending on what resource they want to configure - learn one user interface and you will know how to do this for JBossAS data sources, or OpenSSH daemons, or anything else.