Indeed, I can confirm the issue. The problem comes from the a bit automagic dry-run handling in the @retry decorator (code), that reduces the tries to 1 in dry-run mode but has a bit of heuristic to determine if the current run is dry-run or not.
In this case the problem comes from the fact that the IcingaHosts class doesn't have a self._dry_run property to be checked fo dry-run.

So the course of action here should be:

• modify the `__init__() signature of IcingaHosts` to add the dry_run parameter similarly to many other spicerack classes (example)
• it's possible that the dry_run property will be reported as not used and you might need to add it to the vulture_whitelist file.
• modify the icinga_hosts() method in the Spicerack class to pass the dry_run parameter when instantiating the class (example)
• check if there are other uses of @retry across the codebase in classes that don't have the dry_run parameter to add it to them too if missing
• [stretch/bonus points] look if we can have an easy was to check that in CI so that we can report usages of @retry in places that don't have dry_run awareness.

Missing the two last bulletpoints:

• check if there are other uses of @retry across the codebase in classes that don't have the dry_run parameter to add it to them too if missing
• [stretch/bonus points] look if we can have an easy was to check that in CI so that we can report usages of @retry in places that don't have dry_run awareness.