Remove attempt number from outofproc connection metric

I'm seeing some system metrics that look like this:

`host.startup.outofproc.nodeworker.initialize.attempt0.latency`
`host.startup.outofproc.nodeworker.initialize.attempt1.latency`

For the purposes of aggregation, I want to treat all of these attempts the same way. I'd like to remove the attempt info from the name of the metric. We could maybe put it into the EventMessage column instead.

So I'm suggesting the EventName should be:
`host.startup.outofproc.nodeworker.initialize.latency`