[Monitoring] Share alerting code better

[chrisronline]

Right now, each one of our alerts (cpu usage for example) contains a lot of similar code to the rest of the alerts. Now that we have a few of these under our belt, we need to figure out the right abstraction to use here.

AC:

• Shared threshold alert functionality is moved into base alert
• Server/UI shared elements moved into common folder

Notes:

• We should log a separate ticket to figure out a better abstraction for the shared threshold UI code

[elasticmachine]

Pinging @elastic/stack-monitoring (Team:Monitoring)

[chrisronline]

I'd like to also add we should normalize on adding these two properties to every alert state:

• stackProductUuid
• stackProductName

We should add them to the generic AlertState interface and make them optional because some alerts won't apply to a specific product instance/node.

Or maybe we can create two interfaces, one for alerts that apply to specific instances/nodes and one that does not.

Either way, we need to consolidate so we can leverage more generic flows in the client, such as generically showing the stack product name in places (right now, we can't easily because alerts store them differently)