Currently, when scheduled queries fail (due to error or timeout), users don't receive any notification about it. We should send out an aggregated e-mail to the organization admin with information on failures, which should include for each failure:

1. Which query failed
2. When it failed
3. What is the failure

Notes:

• This feature should be turned off by default and can be turned on with a setting, in order to avoid surprises when upgrading.
• We should default to an hourly e-mail, and the frequency should be configurable.
• If a query failed over threshold times, don't report it.
• If a query failed over 75% of threshold times, mention in the e-mail that this query will soon be unreported.
• threshold should be configurable.