Factor trigger startup failures into health checks

I've had a support case where a site had multiple triggers failing to start due to DNS issues. The failures were permanent, yet we attempted to connect for 6+ days and failed every time.

However, because some other triggers *did* start successfully, the host considered itself in "partial" startup mode. Some functions ran successfully, but the DNS issues impacted other operations as well.

It seems like we could factor this into our Health Checks that we do. If we see triggers not being able to start for some period of time, we consider the host unhealthy and restart the process.