Improve config file watcher resilience

[jaydenthorup]

We now handle Rename events, retry watcher add up to 5 times, and reset retry count after 24 hours of no failures.

[mrlhansen]

I am trying to understand the use case here. You have a scenario where the config file is renamed, and a new file is then eventually renamed to the old name?

[jaydenthorup]

I have a script where the config file is updated every hour based on a nautobot database. The file refresh at times either didn't update or the watcher process crashed out in the middle of the write. Depending on what writer I use for the config, the change is also not caught by the watcher without the rename added.

[mrlhansen]

I will try to have a look, but this feels like a very hacky way of doing it.

[mrlhansen]

I have another suggestion. Would it be useful if the exporter had a /reload endpoint that could be called via a GET request and that would trigger the reload of the configuration file? If you can incorporate that in the workflow.

[mrlhansen]

I added the /reload endpoint in the new release: https://github.com/mrlhansen/idrac_exporter/releases/tag/v2.4.0