Summary Quantiles/Errors

[rmccorm4]

Hi, I had a few questions revolving around value settings for quantiles and their corresponding errors.

I noticed there are no errors or exceptions for setting quantile<0 or quantile>1, and similarly for error<0 or error>1. However, when these values get pushed through upon metric creation, I'm not sure what they're actually interpreting.

I see in the Java client, some of the behavior around zero values is explicitly documented:

• You can set an allowed error of 0, but then the Summary will keep all observations in memory.
• You can track the minimum value with .quantile(0.0, 0.0). This special case will not use additional memory even though the allowed error is 0.
• You can track the maximum value with .quantile(1.0, 0.0). This special case will not use additional memory even though the allowed error is 0.

And values outside of the 0/1 bounds are rejected.

Can you help explain the expected behavior for similar values in the prometheus-cpp implementation?

CC @Romain-Geissler-1A @jupp0r

[dc6c]

Tracking maximum/minimum value is exactly the use case that led me to this issue -- it would be great to see the Quantile(1, 0) case optimized like that!