broadcast: only send bad health notification after consecutive issues

[saxon-milton]

We frequently get a situation where there's a temporary health issue like "noData". This is generally just due to a momentary lapse in bandwidth. We therefore don't care all that much unless it continues. So let's only use email notification if we get consecutive problems.

[scruzin]

This might also be a good use case for sending an email alert to a dedicated email address, such as broadcast-ops@. Less serious alerts could continue to use ops@.