We made an alert to shoot out an e-mail once our number of devices that are live drops by over 20%. We hard coded a new column to store the percentage drop. Works great - the percentage drop hovers between -1% and 1%, but once we experience an outage, it may exceed 20%. The percentage change usually stabilizes after the drop until the issue is resolved, so the next record will be well under 20% - and Redash sends out an OK email which might be interpreted as a resolution. Is there any way to prevent the follow-up alert?