Optional argument to ignore bad configs

[absolutejam]

Feature Request

Proposal:

Argument to allow Telegraf to ignore bad config files instead of throwing a critical exception.

Current behavior:

Currently, a bad telegraf config (Within --config-directory) will kill the Telegraf agent.

Desired behavior:

Option to ignore bad config file and log an message instead.

Use case: [Why is this important (helps with prioritizing requests)]

I didn't see a way to validate configs (Did I miss this), and at the moment, a badly rendered config file from SaltStack causes the entire service to fail. I'd like to ignore bad configs and allow metrics collection to continue instead of having a blackspot in our collection.

On a side note, is there a way to validate configs?

Cheers!

[absolutejam]

Actually, it looks like the --test param can validate the config, but I'd still like this option to be considered 😄

[danielnelson]

I think this would be a good feature, maybe not for the agent section but at least for plugins, and I especially like the idea of an option to validate the config.

Here is a related issue about Telegraf not starting when an ouput connection cannot be made: #3723

[sspaink]

Closing in favor of: #5507
I think adding a way to validate configs is a good idea.