internal/dag: Add support for per-host max connections

[relu]

Support per-host max connections circuit breaker threshold enabling Envoy to enforce a maximum number of connections for each individual Kubernetes service endpoint using a new service-level annotation: projectcontour.io/per-host-max-connections.

Resolves #6015

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (4d48db7) 78.68% compared to head (6d517e5) 78.69%.
Report is 16 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #6016   +/-   ##
=======================================
  Coverage   78.68%   78.69%           
=======================================
  Files         138      138           
  Lines       19675    19681    +6     
=======================================
+ Hits        15481    15487    +6     
  Misses       3890     3890           
  Partials      304      304           

Files	Coverage Δ
internal/annotation/annotations.go	100.00% <100.00%> (ø)
internal/dag/accessors.go	87.39% <100.00%> (+0.05%)	⬆️
internal/dag/dag.go	98.70% <ø> (ø)
internal/envoy/v3/cluster.go	96.36% <100.00%> (+0.04%)	⬆️

[davinci26]

Envoy docs are a bit confusing to me, does host refer to cluster endpoint in this case?

[davinci26]

We need a release note for this.

#6013 is also conflicting. I think we need to be able to set a global default for this.

[davinci26]

otherwise the PR lgtm, thanks for the feature, we might even put it into use actually

[relu]

does host refer to cluster endpoint in this case?

That is what I understand from here.

I think we need to be able to set a global default for this.

That could make sense if one wants to configure a safe default max-connections limit covering all ingress configurations.

Sorry, it seems I forgot to include the release note. I'll add it in a few minutes. Thank you!

[davinci26]

LGTM if you merge before #6013 gets in then you have to do the defaults otherwise I can do it

[github-actions]

The Contour project currently lacks enough contributors to adequately respond to all PRs.

This bot triages PRs according to the following rules:

• After 14d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, the PR is closed

You can:

• Ensure your PR is passing all CI checks. PRs that are fully green are more likely to be reviewed. If you are having trouble with CI checks, reach out to the #contour channel in the Kubernetes Slack workspace.
• Mark this PR as fresh by commenting or pushing a commit
• Close this PR
• Offer to help out with triage

Please send feedback to the #contour channel in the Kubernetes Slack

[davinci26]

@projectcontour/maintainers can you PTAL so we can merge it.

It has some conflicts with #6013 so whoever merges it first need to handle it

[skriss]

@davinci26 thanks for reviewing, I'll take a look this week

[skriss]

thanks @relu, LGTM

[skriss]

@sunjayBhatia will give you a chance to take a look if you want