[Feat]: Consider special alert's attributes when displaying them on cloud.

[MrZammler]

Problem

Not so much as a problem, as more as nice to have and discussion.

Right now we treat all of the alerts mostly the same when it comes to displaying them on the cloud (mostly in the alert drawer, when we draw it over a chart).

We kinda treat the alert's when timestamp, as the moment the "problem" occurred, but that might be misleading.

Some of the alerts specify a long time of metrics on which to apply, i.e. 10min_disk_utilization will check 10 minutes of disk utilization to decide of whether or not to raise it. So it's when timestamp doesn't really say that that time trouble started.

Furthermore, in some cases for critical alerts, we should even consider it's past warning state if we want to point the user to when he should check the state.

Description

Consider the timeframe of metrics an alert calculated.

Importance

nice to have

Value proposition

1. Have a better understanding why the alert was raised.

Proposed implementation

It might be possible for the agent to transmit the range of metrics used to calculate the alert.