fix: update disruption metrics even without candidates

[moko-poi]

fix: update disruption metrics even without candidates

Fixes #2344

Description

The karpenter_nodepools_allowed_disruptions metric was only being updated when disruption candidates existed. This caused the metric to report stale values in several scenarios:

• When no nodes are eligible for disruption
• When schedule-based budgets change (e.g., entering/exiting disruption windows)
• When NodePools are deleted but their metrics remain

Root Cause:
The BuildDisruptionBudgetMapping() function was called after checking if candidates exist. When len(candidates) == 0, the function returned early without updating metrics.

Solution:
Move BuildDisruptionBudgetMapping() before the candidate check. This ensures metrics are updated on every reconcile loop (every 10 seconds), regardless of whether disruption candidates exist.

Impact:

• Metrics now accurately reflect schedule-based budget restrictions in real-time
• Deleted NodePools naturally stop being reported (only active pools are updated)
• No breaking changes - existing behavior with candidates is preserved
• Minimal performance impact - same function calls, just reordered

How was this change tested?

• Code compiles successfully: go build ./pkg/controllers/disruption/...
• No new test failures - existing tests pass as candidate behavior is unchanged
• Manual verification: metric update logic is triggered for all disruption reasons during reconcile, independent of candidate existence
• Logical validation: BuildDisruptionBudgetMapping() is side-effect free except for metrics/events, so moving it earlier is safe

Example Scenario Fixed:

spec:
  disruption:
    budgets:
    - duration: 12h00m0s
      nodes: "0"
      schedule: 0 10 * * mon-fri  # Block disruption 10:00-22:00 weekdays
    - nodes: "2"

Before: Metric shows 2 even during blocked window
After: Metric correctly shows 0 during 10:00-22:00 and 2 outside that window

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

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: moko-poi
Once this PR has been reviewed and has the lgtm label, please assign maciekpytel for approval. For more information see the Code Review Process.

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

Details

Needs approval from an approver in each of these files:

• OWNERS

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

[k8s-ci-robot]

Hi @moko-poi. Thanks for your PR.

I'm waiting for a github.com member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

Details

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

[coveralls]

Pull Request Test Coverage Report for Build 20390985201

Details

• 5 of 5 (100.0%) changed or added relevant lines in 1 file are covered.
• 5 unchanged lines in 2 files lost coverage.
• Overall coverage decreased (-0.005%) to 80.325%

---

Files with Coverage Reduction	New Missed Lines	%
pkg/controllers/disruption/drift.go	2	88.0%
pkg/controllers/provisioning/scheduling/nodeclaim.go	3	89.63%

Totals
Change from base Build 20384548641:	-0.005%
Covered Lines:	11966
Relevant Lines:	14897

---

💛 - Coveralls

[DerekFrank]

I'm not sure this is the right refactor here. Because the metrics are entirely based on cluster state, I think I would prefer if we move the metrics outside of BuildDisruptionBudgetMapping and into the nodepool or node metrics controller. That ensures the metrics will always be emitted no matter the changes to consolidation or disruption.

[moko-poi]

@DerekFrank
You're absolutely right. I've implemented the refactor following your suggestion. I moved the metrics emission out of BuildDisruptionBudgetMapping and into the nodepool metrics controller, which ensures the metrics are always updated regardless of whether disruption candidates exist or consolidation logic changes.

Specifically, I added cluster state and clock dependencies to the nodepool metrics controller and implemented an updateDisruptionBudgetMetrics method. This method is called for each NodePool during the Reconcile loop, counts nodes directly from cluster state, evaluates schedule-based budgets using the clock, and sets the metrics accordingly. Since the controller requeues every 5 minutes, the metrics are updated periodically independent of the disruption controller's execution.

This change makes the metrics logic completely independent from the disruption controller logic, making it more robust. Thank you for the architectural improvement suggestion.

b8227fb

[k8s-ci-robot]

PR needs rebase.

Details

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.