Nate Nuss

A read-through of the graphite [grammar](https://github.com/graphite-project/graphite-web/blob/master/webapp/graphite/render/grammar.py#L83) gives me: ASCII 33 thru 126 (pyparsing `printables`): ``` !"#$%&'()*+,-./ 0123456789 :;?@ ABCDEFGHIJKLMNOPQRSTUVWXYZ [\]^_` abcdefghijklmnopqrstuvwxyz {|}~ ``` While these symbols must be `\` escaped...

highestCurrent is being called but requires an additional 'n' argument which you are unable to pass through groupByNode.

In this specific case a default of 1 makes sense to me. When I test however I find more work is needed because highestCurrent is a filter (returns original series...

Something like this could work but I'm unhappy enough with it (always walking all the map, append(results, r...) when r might now have >1, ... ) that I'm just putting...

`func EvalExpr( e *expr, from, until int32, values map[MetricRequest][]*MetricData ) ([]*MetricData, error)` `values` looks like: ``` values["metric1.*.total"] = [ &{"metric1.A.total", ...series data...}, &{"metric1.B.total", ...series data...}, &{"metric1.C.total", ...series data...}, &... more...

> If resetEnd is False, will instead draw full range including future time. My read of the doc is the opposite. That we always behave like `resetEnd=false`. I'm also experiencing...

@JaderDias also pointed out that my proposal in #222 would be sugar around what can already be done: - set the mask metric points to FloatMax (rather than 1) -...

(also here) The history around why carbonzipper went with the naming scheme it did: https://github.com/go-graphite/carbonzipper/pull/18

This is nearly 2 years old - think we should still look at it?

Re: 2) Graphite has moved to a single-phase `/render` request to the backing stores. If we also make that change as well as carbonserver support for multiple target: `/render/?from=147...&to=147...&format=...&target=metric.a&target=metric.b&target=...` ;...