I have built a CRUD app with Blazor that I ported to Blazor Serverside and now Razor Components (with ASP.NET Core 3.0 preview2).
Since going serverside, I find intermittently that the app freezes and stops responding to user input. I have not yet had a chance to look into what is going on, but wonder if anybody else has experienced this and if it is a bug. My observations and thoughts are:
I will attempt to diagnose this next time it happens, but I am very curious to know whether this has been experienced by anybody else. I can grant access to the app to a team member if asked, although there is a commercial sensitivity to this.
Many thanks, Simon
I managed to capture information from the console (using Edge) when I got a crash. Sorry for having redacted the website's URL.

Here is another mode of failure that I find from time to time, but mostly after the "Log In" button is pressed on my login splash screen at startup.

Looking into this I find the following:
My takeaway is that Razor Components might need to be more resilient so that it avoids losing its connection arising from App bugs. I'll leave it to the ASP.NET team to decide when to close this issue.
Razor Components is still great stuff though. Simon
Thanks for contacting us, @simonziegler.
- Razor Components however loses its SignalR connection when one of these bugs gets hit, which is why I wondered if the bug was actually in Razor Components
@SteveSandersonMS, @javiercn what do we plan to do about this?
/cc @danroth27
The work that @javiercn and @pranavkm are doing at the moment change the unhandled exception behavior, so it should log the errors and as far as I know will no longer drop the whole circuit (connection). Longer term we have a whole set of work planned to deal with automatic reconnection for the case where the SignalR connection drops due to network issues.
Closing this because this work is already tracked elsewhere. Hope that sounds reasonable.
Thanks All. I'm happy to know that you are aware of this and have a plan. That's all I need.
Most helpful comment
The work that @javiercn and @pranavkm are doing at the moment change the unhandled exception behavior, so it should log the errors and as far as I know will no longer drop the whole circuit (connection). Longer term we have a whole set of work planned to deal with automatic reconnection for the case where the SignalR connection drops due to network issues.
Closing this because this work is already tracked elsewhere. Hope that sounds reasonable.