Performance job for NFTables

[aroradaman]

There were discussions in NFTables KEP regarding performance comparison with IPTables proxy.

We are looking out for a perf job for NFTables proxy but aren't sure about the size/scale of the cluster, load on the cluster, frequency of the job, and other things. Along with comparing the perf with IPTables, this will also help us compare perf with possible future optimizations within NFTables proxy.

ci-kubernetes-e2e-gce-scale-performance tracks NetworkLatency and NetworkProgrammingLatency for IPTables proxy at scale, maybe we need similar metrics at a low scale?

Opening up this issue for discussion.

/sig network
/sig scalability

/cc. @danwinship @aojea @thockin

[aroradaman]

/cc. @ameukam

[aroradaman]

/sig network
/sig scalability

[aojea]

@npinaeva and I are working on this, no need to go to scale in a first iteration, we want to validate first the improvements of nftables over the well known iptables problems:, that are:

• programming latency: iptables-save and restore with large numbers of rules
• latency degradation vs number of rules caused by the linear search on the iptables rules

[k8s-triage-robot]

The Kubernetes project currently lacks enough contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue as fresh with /remove-lifecycle stale
• Close this issue with /close
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

[npinaeva]

/remove-lifecycle stale

[aojea]

#32431