Align tests and schemas

[hamistao]

As part of a personal Goal, I experimented with using AI to identify discrepancies between our automated tests and their schemas and this PR is a result of trying to fix the identified discrepancies.

The more convoluted fixes here would be:

• The helm install commands from TestAmbientInstallation and TestGatewayStandaloneInstallation schemas were updated to match the one used on the tests.
• TestLoadBalancer and TestLoadBalanceScaleAndUpgrade were updated to test the connectivity to the actual Load Balancers instead of the Node port

[Copilot]

Pull request overview

This PR aims to align PIT schemas with existing automated test behavior and to adjust load balancer connectivity tests to validate connectivity against the actual LoadBalancer endpoint rather than NodePort, reducing schema/test drift and improving real-world coverage.

Changes:

• Refactors workload deployment test table setup to remove redundant per-test create function wiring.
• Updates PIT schema steps/commands/positions across multiple suites to better match test execution.
• Updates LoadBalancer tests to validate connectivity via the LoadBalancer service address and introduces a new helper for that validation.

Reviewed changes

Copilot reviewed 10 out of 10 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
validation/workloads/workload_test.go	Simplifies deployment test case struct by removing redundant create function field.
validation/workloads/schemas/pit_schemas.yaml	Adjusts workload PIT schema formatting and scale-step definitions to match test behavior.
validation/upgrade/schemas/pit_schemas.yaml	Updates PIT schema formatting and step position numbering for upgrade workload cases.
validation/networking/connectivity/port_test.go	Switches LoadBalancer connectivity checks from NodePort to LoadBalancer service endpoint and simplifies test structure.
validation/longhorn/schemas/pit_schemas.yaml	Updates Longhorn PIT schema steps to reflect revised replica/scale expectations (with a noted grammar fix needed).
validation/fleet/schemas/pit_schemas.yaml	Updates expected namespace wording in Fleet PIT schema to match test expectation.
validation/charts/schemas/pit_schemas.yaml	Updates chart PIT schema commands/details (alerting namespace, neuvector chart list) to match tests.
validation/charts/appco/schemas/pit_schemas.yaml	Updates Istio AppCo standalone install schema command (currently diverges from test behavior regarding imagePullSecrets).
actions/workloads/deployment/verify.go	Fixes replica update logging to reflect the computed target replica count.
actions/networking/verify.go	Adds LoadBalancer connectivity helper (currently contains correctness/compilation issues that need fixes).

[slickwarren]

I think we still need changes here; can you send a link to a test run using this?

Lastly can you confirm all callers of the changed signatures are updated:

VerifyClusterConnectivity changed its parameter from path string to port int. Please confirm there are no other callers of VerifyClusterConnectivity or VerifyNodePortConnectivity elsewhere in the repo that still use the old signature.

[Copilot]

Pull request overview

Copilot reviewed 9 out of 10 changed files in this pull request and generated no new comments.

Comments suppressed due to low confidence (7)

actions/networking/verify.go:155

• Go error strings should be lower-case. This error message currently starts with a capital letter.

		return fmt.Errorf("Curl result %s doesn't include the workload name %s", log, workloadName)

actions/networking/verify.go:190

• This assumes the LoadBalancer ingress always has an IP (Status.LoadBalancer.Ingress[0].IP). On some providers it is only a hostname, so ip can be empty and the curl target becomes invalid. Also, the newly introduced error strings should be lower-case to follow Go conventions.

	port := k8sService.Spec.Ports[0].Port
	ip := k8sService.Status.LoadBalancer.Ingress[0].IP
	logrus.Infof("Testing connectivity with load balancer %s by curling %s:%d/name.html", k8sService.Name, ip, port)

validation/networking/connectivity/port_test.go:443

• Spelling/terminology: this skip message refers to "Load Balance" but the test is for a Kubernetes Service of type "LoadBalancer".

		p.T().Skip("Load Balance test requires access to cloud provider.")

actions/networking/verify.go:121

• When ip is empty, the code overwrites the ip parameter inside the loop. That makes subsequent iterations reuse the first resolved node IP, which contradicts the function comment (“Each node's own ip is used as default”) and can make NodePort checks flaky in environments where a node’s external IP isn’t routable from other worker nodes. Use a per-iteration targetIP instead of mutating ip.

		if ip == "" {
			newNode := &corev1.Node{}
			err = v1.ConvertToK8sType(machine.JSONResp, newNode)
			if err != nil {
				return fmt.Errorf("failed to convert node %s: %w", machine.Name, err)

actions/networking/verify.go:144

• Go error strings should be lower-case and not start with a capital letter. This returned error message should be lower-cased for consistency with the rest of the codebase (e.g. "no worker nodes found").

This issue also appears on line 155 of the same file.

	return fmt.Errorf("Unable to connect to %s:%d/name.html from any worker node", ip, port)

actions/networking/verify.go:162

• The function header comment is incomplete ("This includes") and doesn't add useful information. Please either remove the fragment or replace it with a complete description.

// VerifyLoadBalancerConnectivity verifies that the Load Balancer service is accessible by curling its IP:port.
// This includes

validation/networking/connectivity/port_test.go:198

• Spelling/terminology: this skip message refers to "Load Balance" but the test is for a Kubernetes Service of type "LoadBalancer".

This issue also appears on line 443 of the same file.

		p.T().Skip("Load Balance test requires access to cloud provider.")