Hi, we have a lot of Service Bus namespaces that we are monitoring and we have set up metric events around these. I have been observing the metric for a while now and have seen delays up to and around 10 minutes for metrics to come in.
This is causing some weird behaviours for our metric events since they are defined on a sliding window of 10. I'm wondering if there's a way to scale the polling of this metric accordingly?
On a slightly different topic, if "alert on missing data" is disabled are missing data points considered as "good" samples? I.e if we have no data for 10+ minutes would it mean a previously opened metric event will close?