fix(prometheus): flaky metrics server

[dnut]

Problems

Prometheus does not consistently get metrics from our metrics endpoint. The metrics are intermittent.

Also, the server often crashes due to errors. In #590 I added a wrapper to catch these errors and restart the server. But sometimes an error occurs that is handled, a server restart is attempted, but the server never responds to any http requests any more, so the metrics become completely inaccessible until sig is completely restarted. Usually this is caused by error.BrokenPipe

Solution

Revert "fix(prometheus): remove httpz again and fix prometheus metrics (#555)"

This reverts commit 3502333.

I understand the desire to eliminate unnecessary dependencies. In the grand scheme of things, eliminating httpz is not that important. Having a working metrics endpoint is critical though. We have a solution that works with httpz. We should test more thoroughly before replacing it.

[0xNineteen]

if our zig server implementation is flaky rn we should keep this in mind when were using it for rpc (and if we should take a different approach there too)

[codecov]

Codecov Report

Attention: Patch coverage is 0% with 8 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/prometheus/http.zig	0.00%	8 Missing ⚠️

Files with missing lines	Coverage Δ
src/prometheus/http.zig	0.00% <0.00%> (ø)

🚀 New features to boost your workflow:

• ❄ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.