prometheus metric prefix for the router

[garypen]

I noticed in the grafana metric explorer, that most of the metrics available are prefixed with a name representing the executable/service. For example:

jaeger_http_<blah...>
kubelet_http_<blah...>
alertmanager_http_<blah...>

The router doesn't do this. router metrics look like:

http_<blah...>

Now, it's possible (after choosing a metric) to narrow down the search by other labels (such as service or container), so it's definitely possible to find the required router metrics but I wonder if we should adopt this prefix naming convention?