perf: track empty domains separately in topology

[tzneal]

Fixes #N/A

Description

Tracking empty domains separately allows for more efficient anti-affinity checks when there are no empty domains. This is a frequent operation when doing a large scale up of a deployment with self anti-affinity.

Using benchmarks with anti-affinity pods only, scheduling performance increases greatly:

name               old pods/sec  new pods/sec  delta
Scheduling1-12        71.4 ±62%    304.0 ±13%  +326.11%  (p=0.008 n=5+5)
Scheduling50-12       49.2 ±12%    295.2 ± 3%  +499.63%  (p=0.008 n=5+5)
Scheduling250-12       236 ± 4%      455 ± 3%   +92.83%  (p=0.008 n=5+5)
Scheduling500-12       226 ± 7%      364 ± 1%   +61.51%  (p=0.008 n=5+5)
Scheduling1000-12     74.4 ± 4%    349.6 ± 1%  +370.05%  (p=0.008 n=5+5)
Scheduling2000-12     21.7 ± 1%    186.8 ± 1%  +759.80%  (p=0.016 n=5+4)

How was this change tested?

Unit tests

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

[jonathan-innis]

/lgtm
/approve

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: jonathan-innis, tzneal

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [jonathan-innis,tzneal]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[coveralls]

Pull Request Test Coverage Report for Build 7650057606

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.02%) to 80.37%

---

Totals
Change from base Build 7646651006:	0.02%
Covered Lines:	7738
Relevant Lines:	9628

---

💛 - Coveralls