Improve API for handling custom metric dimensions

[ysbakker]

Is your feature request related to a problem?
Right now, if we want to write an abstraction for creating metrics with an n number of custom dimensions, we have to do it like this:

public void LogMetric(string metricNamespace, string metricName, double value, IDictionary<string, string> dimensions)
{
    var id = new MetricIdentifier(metricNamespace, metricName, dimensions.Keys.ToList());
    var metric = _telemetryClient.GetMetric(id);
    if (metric.TryGetDataSeries(out var series, true, dimensions.Values.ToArray()))
    {
        series.TrackValue(value);
    }
}

Metric.TrackValue has a bunch of overloads so it supports up to 10 dimensions. There is no overload for providing a collection of dimensions. Internally, Metric.TryGetDataSeries is called and this does have an overload for supplying n dimensions.

Describe the solution you'd like.
Create a Metric.TrackValue overload that accepts a collection of dimensions. This should be fairly straight-forward as dimensionality validations are already in place. So supplying too many / too little dimension values should already throw a validation error.

Additionally, a Metric.GetValue overload that accepts n dimensions would be nice. As of right now, there are only overloads for up to 4 dimensions. So metrics with more than 4 dimensions have to be created using a custom MetricIdentifier.

Describe alternatives you've considered.
Using the MetricIdentifier and Metric.TryGetDataSeries methods manually. But in my opinion, this fits in the existing API.

Additional context.

none.