Feature Request: Trigger alert based on result of previous alert

This should address below scenario.

Say we have configured **tick** script to alert(email for ex) if node goes down. There might be two reasons for node going down. One, telegraf agent is down. Two, node itself is down. In this case we want to evaluate what is the reason for node going down.

Here, we can configure ` .exec` and run some script to analyze the node failure and then return some result back to Kapacitor and then Kapacitor should alert based on the result received.

This would be a cool feature if implemented.

Start by reviewing Kapacitor's .exec behavior and how alert results are currently evaluated. Trace how a script result could be returned to an alert rule, then define tests for distinguishing a down Telegraf agent from a down node and triggering the corresponding alert. Done means the requested result-based alert flow is specified and covered by tests.

Written by the indexing model from the issue text.