Functions with QueueTrigger misconfigured make the host restart over and over

If there is a QueueTrigger in an AzureFunction that does not work (eg Connection does just not exist), this will make the host restart. I think this is not a good behavior, as it results in the following user experience:

- Functions App restarts and restarts and restarts and restarts...
- Log Stream works on Windows only (at least for me). Therefore, diagnosing this can be very hard on Linux. Maybe it's possible to diagnose this with Azure tools, but I was not able to.
- App Insights does not retrieve the error
- Some requests work, but they are very slow even for trivial stuff and they only work if you are in luck and the host is not restarting already.

I propose to just ignore not-working queue triggers and log them to app insights (if configured) and to Log Stream.

If I recall correctly I had similar error behavior on other mis-config as well; maybe it's a general problem that error handling on startup is not very forgiving.

Of course it's more of an user error, but the UX is surely not optimal. And as it's Azure I think it should be awesome :)

Start by reproducing a QueueTrigger configured with a nonexistent connection and observe the host restart and available logs. Done means the misconfigured trigger no longer causes repeated host restarts and the failure is visible through the requested diagnostics, including Application Insights when configured and Log Stream.

Written by the indexing model from the issue text.