Best practice for zone config?

Is this still recommended as best practice? As of this and if the director is used, I’d assume a Kickstart Wizard run is then required (to import these settings for creating host object in these zones). Am I right?

If so, how can I “repair” my setup where I started without according zone and endpoint information within conf files but directly in the director?

If not, zones and endpoints shall be configured in the director only (except for each node itself)?


the blog post doesn’t mention the Director, more relatively the Puppet module. I for myself would manage the master and satellite zone configuration outside of the Director, and only import that via kickstart wizard. That allows to use zones/endpoints as external objects and assign the cluster zone settings wherever needed.

The agent/client configuration can be fully managed by the Director, the setup routines depend on what you like, either the provided scripts inside the Director, or you’ll go with the CLI tools / Windows setup wizard provided with the Icinga 2 distributed monitoring chapter.

One thing to note: Even if there is the Infrastructure section in the Director which allows you to create zones/endpoints on your own, this will become a chicken-egg problem with deployments and connection required before the deployment actually can trigger a sync. Do not manage master/satellite zones inside the Director is my advice here.


1 Like