First and foremost, it monitors the real world, not the version control system. It sees things as they are, not as they're supposed to be. It detects manual, unversioned changes. It detects changes from automation, or from intrusion. It detects things that can't be versioned, like reboots or cert expirations.

Second, it is minimally invasive. You don't have to change your process in order to use it.

Third, it shows relationships between systems. An application isn't just source code. It's compiled binaries, OS patches, user security, databases, firewalls, message queues, apis, and more. Here's a common failure pattern - System A changes, System B breaks. System A might actually be perfectly fine and correct, and B as well, but the relationship between A and B is broken. Change the database schema, and the app can no longer read it. Change the dns, and the UI can no longer find the API. For any decently complex application, the potential sources of failure are in the thousands, and potential interactions effectively infinite. Now scale this up to the enterprise, to organizations so large that your app can be broken by departments you didn't know existed...

Does it detect when you have a busted marketing site?

https://congruence.io

Thanks. It's a known problem... we're currently in product work and haven't been dealing with marketing at all. It's just a cert problem - http://congruence.io still works.

And yes, once it's working, it will be able to detect such issues.

And the certificate is now fixed. Try it, https://congruence.io is working again!