Phoenix is configured so that in prod configuration no errors are shown in the response, you wouldn’t want someone getting logs from your application when something goes wrong for a deployed instance. That obviously can be changed from configuration, look what config phoenix is using for logging in dev.exs and update it in prod.exs or runtime.exs.
Errors with code status 500 are always related to server errors, this means something unexpected went wrong on the server. In this case, this error is returned because the process serving the request crashed. This functionality should not be tested, as this is considered a bug that should be fixed.
For errors that come from controllers, I strongly recommend using action_fallback/1 instead of mechanism of catching errors.
This likely works in dev, because phoenix enables Plug.Debugger in development. There can only ever be one error handler though, so your Plug.ErrorHandler might not even run.
I am also trying to setup telemetry, error-logging in my app.
Could you possibly direct me to some good resources around this. I am new to Elixir/Phoenix & getting a bit confused around setting things up properly. I am reading Elixir-In-Action & Phoenix-In-Action books on the side to understand the ecosystem better but need to ship too.
For observability & monitoring (esp Sentry for now) I found the below resources. Would love to have something which is more comprehensive & which explains the internals also… Also something around monitoring sockets, channels would be great.