I intentionally picked a new one that will need queries to be modified if we want to use the metrics from the older packages. I did it so that we can trust the metrics more, because right now when querying for GamerHash group we'd get metrics from both polygon and holesky (or whatever the testnet is). If we want to query both, we can do it, but need to opt into it by adding or to the query. I prefer such explicity.
I checked that GamerHash dashboards on grafana don't use group label - this PR won't break them (though perhaps after deploying a new package we should update the dashboards)
edit: my bad, the dashboards use it, but are also already ready for selecting other names
I thought a little whether the name should contain currency too so that it is more similar to the metrics we collect from our requestor, but I couldn't find it anywhere in the code, so I think we should be fine without it.
I didn't get any provider logs that pushing metrics on yagna exit would help with, so I'll leave that for another time.
With regard to the
grouplabel:GamerHashgroup we'd get metrics from both polygon and holesky (or whatever the testnet is). If we want to query both, we can do it, but need to opt into it by addingorto the query. I prefer such explicity.I checked that GamerHash dashboards on grafana don't usegrouplabel - this PR won't break them (though perhaps after deploying a new package we should update the dashboards)I didn't get any provider logs that pushing metrics on
yagnaexit would help with, so I'll leave that for another time.