test: expand LRP test to include lifecycle events

[karina-ranadive]

Reason for Change:
Test Steps:

1. Initial Validation - Performs an initial DNS test to verify LRP is working correctly with metrics validation
2. Cilium Command Validation - Uses Cilium CLI commands to validate that LRP is properly configured in the Cilium dataplane, checking both cilium lrp list and cilium service list outputs
3. Pod Restart Testing - Restarts client pods and verifies that LRP functionality persists after pod restarts, ensuring the policy survives pod lifecycle changes
4. Post-Restart Validation - Re-runs DNS tests and Cilium validation to confirm LRP still works after client pod restarts
5. Resource Recreation Testing:

• Deleting client daemonsets and LRP resources
• Recreating the LRP policy
• Restarting node-local-dns pods to pick up the new configuration
• Recreating client daemonsets

6. Port Forward Re-establishment - Sets up a new port forward to the newly restarted node-local-dns pod (using a different local port to avoid conflicts)
7. Final Metrics Validation - Validates that DNS requests are properly redirected to the new node-local-dns pod through metrics monitoring
8. Final Cilium Validation - Performs a final check using Cilium commands to ensure LRP is still active after all the resource recreation

Purpose: This comprehensive lifecycle test ensures that Local Redirect Policy functionality is robust and survives various operational scenarios including pod restarts, resource deletion/recreation, and service restarts - validating both functional behavior through DNS metrics and dataplane configuration through Cilium CLI inspection.

Issue Fixed:

Requirements:

• uses conventional commit messages
• includes documentation
• adds unit tests
• relevant PR labels added

Notes:

[Copilot]

Pull Request Overview

Adds negative Local Redirect Policy (LRP) test scenarios and parameterizes Prometheus endpoint usage in existing LRP tests.

• Extends testLRPCase to accept a Prometheus address for flexible metric validation.
• Introduces comprehensive negative and resilience LRP tests (resource recreation, pod restarts, cilium validation).
• Adds helper functions for validating Cilium LRP state and recreating test resources.

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.

File	Description
test/integration/lrp/lrp_test.go	Adds negative LRP test flow, Prometheus address parameter, and Cilium/LRP validation helpers.
test/integration/lrp/lrp_fqdn_test.go	Updates calls to testLRPCase to include the new Prometheus address parameter.

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

The testify assertions use formatted messages with placeholders but require.Greater and require.Equal do not perform format substitution; use require.Greaterf / require.Equalf or pre-format the message with fmt.Sprintf to ensure values appear in failure output.

Suggested change

	require.Equal(t, afterValue, beforeValue, "dns metric count increased after command - before: %.0f, after: %.0f", beforeValue, afterValue)
	require.Equalf(t, afterValue, beforeValue, "dns metric count increased after command - before: %.0f, after: %.0f", beforeValue, afterValue)

[Copilot]

Fixed sleep introduces unnecessary delay and potential flakiness; replace with a poll-based wait (e.g., repeatedly checking for resource absence/recreation with a timeout) to reduce test duration and improve reliability.

Suggested change

	// Wait for deletion to complete
	time.Sleep(10 * time.Second)
	// Wait for deletion to complete (poll for absence of client DaemonSet and LRP)
	retry.DoWithTimeout(ctx, "wait for client DaemonSet deletion", 30*time.Second, func(ctx context.Context) (bool, error) {
	_, err := dsClient.Get(ctx, clientDS.Name, metav1.GetOptions{})
	if err != nil {
	// DaemonSet not found
	return true, nil
	}
	// Still exists
	return false, nil
	})
	retry.DoWithTimeout(ctx, "wait for LRP deletion", 30*time.Second, func(ctx context.Context) (bool, error) {
	_, err := lrpClient.Get(ctx, lrp.Name, metav1.GetOptions{})
	if err != nil {
	// LRP not found
	return true, nil
	}
	// Still exists
	return false, nil
	})

[Copilot]

Pull Request Overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.

Comments suppressed due to low confidence (1)

test/integration/lrp/lrp_test.go:517

• The rand.Seed() call is made every time TakeOne() is invoked, which can cause issues in concurrent tests or when called multiple times in quick succession (as the seed would be very similar). Additionally, rand.Seed() has been deprecated since Go 1.20 in favor of using the global generator which is automatically seeded.

Recommendation: Remove the rand.Seed() call and use the global random generator directly, which is automatically seeded in Go 1.20+:

func TakeOne[T any](slice []T) T {
    if len(slice) == 0 {
        var zero T
        return zero
    }
    return slice[rand.Intn(len(slice))]
}

Alternatively, if you need better randomness, use crypto/rand or math/rand/v2 for Go 1.22+.

// TakeOne takes one item from the slice randomly; if empty, it returns the empty value for the type
// Use in testing only
func TakeOne[T any](slice []T) T {
	if len(slice) == 0 {
		var zero T
		return zero
	}
	rand.Seed(uint64(time.Now().UnixNano()))

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[jpayne3506]

Is it reasonable to retry for 15 minutes (900seconds)? How long does this process normally take?

If it is what I expect (less than 10 seconds) then we can leverage Delete* consts. Anything more than that is excessive.