Closed ryzngard closed 1 week ago
I'm wondering if we can remove this complexity altogether. I don't currently see a need to separate out the pools. @ToddGrun @dibarbet is there something I'm missing here? It looks like each https://github.com/ToddGrun/roslyn/blob/2cd8f0b6930a69f973454ecb8daf474c2d68d083/src/VisualStudio/Core/Def/Telemetry/Shared/AggregatingTelemetryLog.cs#L20 contains a set of telemetry events to be reported by the same name but with different metrics. Is there a logical reason to separate that so there's multiple pools?
I'm wondering if we can remove this complexity altogether. I don't currently see a need to separate out the pools. @ToddGrun @dibarbet is there something I'm missing here? It looks like each https://github.com/ToddGrun/roslyn/blob/2cd8f0b6930a69f973454ecb8daf474c2d68d083/src/VisualStudio/Core/Def/Telemetry/Shared/AggregatingTelemetryLog.cs#L20 contains a set of telemetry events to be reported by the same name but with different metrics. Is there a logical reason to separate that so there's multiple pools?
Not sure if you saw, but david did some cleanup about a month ago including the file you are referencing in this PR
When first introduced there was an error in assumption of how histogram telemetry worked. This fixes this by correctly reporting a unique histogram even per lsp method name rather than the first one that comes in (which, as it turns out, was always initialize)
Also cleaned up the wording so it was a little easier to understand.