Thanks for the PR.

I agree that silent failure's probably a safe default in a lot of cases, but it also has the likelihood of burying errors and making it more difficult to detect failures.

Any project without process monitoring will be reliant on manual testing of functionality, which will be much more difficult with silent SSR failures. For example, if the browser's cached assets, the time before the client-side code mounting may be imperceptible, even for someone looking for failure.

Can you change the default value of the setting from not Debug to False?

Also, do you mind expanding on the setting in the README? Something like

In some production scenarios it's not ideal to rely on the render server ability to serve a request. If this setting is set to True, the library will silently ignore render server errors and return empty strings as the markup.

I would suggest still throwing the exception, but catching it when the setting is on and logging it. Like what I did with FAIL_SAFE setting on react-render.

That way you still get an error in your logs, and you can still inspect the stack trace in something like sentry.

Thanks for the reply guys!

And both of you are totally right... this was my quick fire response to handling some time out issues (which I should find the source of).

But I'm happy it brought up the discussion. I think the outcome would be adding an example to the README demonstrating how you could avoid hard fails in production. When I put together a solution I will share.