fix(metrics): use f64::MAX as histogram bound to display +Inf properly

[flaneur2020]

I hereby agree to the terms of the CLA available at: https://docs.databend.com/dev/policies/cla/

Summary

on rendering the prometheus metrics, prometheus-client checks the upper bound with f64::MAX, and renders it as +Inf, which is a marker in calculating histogram_quantile on monitoring.

if +Inf is not displayed properly, it seems that histogram_quantile will return nothing.

the source look likes this:

            self.write_prefix_name_unit()?;
            self.write_suffix("bucket")?;

            if *upper_bound == f64::MAX {
                self.encode_labels(Some(&[("le", "+Inf")]))?;
            } else {
                self.encode_labels(Some(&[("le", *upper_bound)]))?;
            }

            self.writer.write_str(" ")?;
            self.writer
                .write_str(itoa::Buffer::new().format(cummulative))?;

to display +Inf properly, we need let the upper bound with f64::MAX.

Tests

• Unit test

Type of change

• Bug Fix (non-breaking change which fixes an issue)

---

This change is 

[github-actions]

Docker Image for PR

• tag: pr-16701-e2f19db-1730194908

note: this image tag is only available for internal use,
please check the internal doc for more details.

[zhang2014]

Cloud test passed