Hi All,
While I know this can be done in DQL - here's a challenge for Managed clients requirement:
a customer wants to create an SLO on a monthly basis. Since the targets and calculations are being added everyday, one Day's 10% of error burn rate may end up eventually as a 1% for the whole month, hence creating false alerts. to try and overcome this, the idea is to try and divide the number of errors for the current month by the average of a monthly total request count for the past 6 months. This practically means that the same metric selector has a divisor that looks at the current month timeframe and a dividend that is actually an average of the total amount for the past 6 months - simply put - two different time frames.
Can this be done somehow? I couldn't figure out a way.
