I had a very similar experience with pi hole. With DNS ad blockers, blocking happens upstream and via a mechanism that causes it to always looks like an error.

I had a very similar experience with pi hole. With DNS ad blockers, blocking happens upstream and via a mechanism that causes it to always looks like an error. So you lose your ability to both troubleshoot the problem and to visualize its behavior in the same context as the application it broke. I'm of the opinion that MITM products are a poor fit for controlling endpoint behavior. Ad blockers, content blockers, security scanners, etc., all tend to give better and clearer results when they operate on the same machine they're affecting.