Skip to content

Calibrate whole-surface migration effort - #147

Merged
laclance merged 1 commit into
devfrom
fix/whole-surface-migration-effort
Oct 7, 2026
Merged

laclance merged 1 commit into
devfrom
fix/whole-surface-migration-effort

Conversation

@laclance

@laclance laclance commented Oct 7, 2026

Copy link
Copy Markdown
Owner

Summary

Fix effort underestimation for broad interface migrations such as aws/amazon-vpc-cni-k8s#3906.

The issue asks to migrate iptables handling to the nftables API for all rules, with Prerouting already migrated and remaining Postrouting/mangle paths still using iptables-nft. The existing estimator classified it as 6–12h solely from the enhancement signal.

This adds a narrow large-effort signal for explicit whole-surface migrations: migration/replacement/conversion language combined with all, every, or remaining rules/paths/calls/uses.

Regression coverage

  • AWS VPC CNI #3906-shaped input estimates 1d+
  • reason is whole-surface interface migration
  • a one-rule API migration remains 6–12h

Scope

Only:

  • opportunity_scout/scoring.py
  • tests/test_effort_preferences.py

No readiness changes, private-instance changes, or release changes.

@laclance
laclance marked this pull request as ready for review October 7, 2026 15:41
@laclance
laclance merged commit e2ef3a3 into dev Oct 7, 2026
4 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant