Add the provider Checker contract and the kubernetes and GCP checks

Checker is an optional interface next to Provider: Check returns one verdict per target and never fails as a whole. The kubernetes provider lists at most one managed Pod cluster-wide (the verb and scope GC relies on); the GCP provider fetches a single page of at most one managed instance through a new Probe seam method, since the SDK treats MaxResults as a page size and the existing AggregatedList drains every page. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-09-11 20:08:21 +00:00
parent 1c773c24ad
commit 1962365f44
9 changed files with 511 additions and 5 deletions
@@ -0,0 +1,53 @@
+package kubernetes
+
+import (
+	"context"
+	"errors"
+	"time"
+
+	corev1 "k8s.io/api/core/v1"
+	apierrors "k8s.io/apimachinery/pkg/api/errors"
+	"sigs.k8s.io/controller-runtime/pkg/client"
+
+	"gitea.home.hrajfrisbee.cz/kacerr/egress-proxies-operator/internal/provider"
+)
+
+var _ provider.Checker = (*Provider)(nil)
+
+// Check lists at most one managed Pod across the cluster: the same client,
+// verb and scope ListByTag and orphan GC rely on, so a pass means the API
+// server answers and the operator's RBAC still covers them. It does not
+// prove create or delete rights.
+func (p *Provider) Check(ctx context.Context) []provider.TargetStatus {
+	start := time.Now()
+	var pods corev1.PodList
+	err := p.client.List(ctx, &pods, client.MatchingLabels{
+		provider.LabelManaged: provider.LabelManagedYes,
+	}, client.Limit(1))
+	ts := provider.TargetStatus{Name: "apiserver", Address: p.apiHost, Latency: time.Since(start)}
+	if err != nil {
+		ts.Reason = checkReason(ctx, err)
+		ts.Message = err.Error()
+		return []provider.TargetStatus{ts}
+	}
+	ts.Ready = true
+	ts.Reason = provider.ReasonReachable
+	return []provider.TargetStatus{ts}
+}
+
+func checkReason(ctx context.Context, err error) string {
+	var status apierrors.APIStatus
+	switch {
+	case ctx.Err() != nil, errors.Is(err, context.DeadlineExceeded),
+		apierrors.IsTimeout(err), apierrors.IsServerTimeout(err):
+		return provider.ReasonTimeout
+	case apierrors.IsUnauthorized(err):
+		return provider.ReasonUnauthorized
+	case apierrors.IsForbidden(err):
+		return provider.ReasonForbidden
+	case errors.As(err, &status):
+		return provider.ReasonAPIError
+	default:
+		return provider.ReasonUnreachable
+	}
+}
@@ -0,0 +1,114 @@
+package kubernetes
+
+import (
+	"context"
+	"errors"
+	"net/url"
+	"syscall"
+	"testing"
+
+	corev1 "k8s.io/api/core/v1"
+	apierrors "k8s.io/apimachinery/pkg/api/errors"
+	"k8s.io/apimachinery/pkg/runtime"
+	"k8s.io/apimachinery/pkg/runtime/schema"
+	"sigs.k8s.io/controller-runtime/pkg/client"
+	"sigs.k8s.io/controller-runtime/pkg/client/fake"
+	"sigs.k8s.io/controller-runtime/pkg/client/interceptor"
+
+	"gitea.home.hrajfrisbee.cz/kacerr/egress-proxies-operator/internal/provider"
+)
+
+type listFunc func(ctx context.Context, c client.WithWatch, list client.ObjectList, opts ...client.ListOption) error
+
+func newCheckProvider(list listFunc) *Provider {
+	scheme := runtime.NewScheme()
+	if err := corev1.AddToScheme(scheme); err != nil {
+		panic(err)
+	}
+	c := fake.NewClientBuilder().WithScheme(scheme).WithInterceptorFuncs(interceptor.Funcs{List: list}).Build()
+	p := newWithClient(c, provider.ProviderConfig{Name: "kubernetes"})
+	p.apiHost = "https://10.96.0.1:443"
+	return p
+}
+
+func TestCheck_reachable(t *testing.T) {
+	t.Parallel()
+	p := newTestProvider(testPod("ns", "a", "uid-a", corev1.PodRunning, "10.0.0.1"))
+	p.apiHost = "https://10.96.0.1:443"
+
+	got := p.Check(context.Background())
+
+	if len(got) != 1 {
+		t.Fatalf("Check() returned %d targets, want 1", len(got))
+	}
+	ts := got[0]
+	if !ts.Ready || ts.Reason != provider.ReasonReachable || ts.Message != "" {
+		t.Errorf("Check() = %+v, want ready and Reachable", ts)
+	}
+	if ts.Name != "apiserver" || ts.Address != "https://10.96.0.1:443" {
+		t.Errorf("target = %q at %q, want apiserver at https://10.96.0.1:443", ts.Name, ts.Address)
+	}
+}
+
+func TestCheck_listOptions(t *testing.T) {
+	t.Parallel()
+	var got client.ListOptions
+	p := newCheckProvider(func(ctx context.Context, c client.WithWatch, list client.ObjectList, opts ...client.ListOption) error {
+		got.ApplyOptions(opts)
+		return c.List(ctx, list, opts...)
+	})
+
+	p.Check(context.Background())
+
+	if got.Limit != 1 {
+		t.Errorf("Limit = %d, want 1", got.Limit)
+	}
+	if got.Namespace != "" {
+		t.Errorf("Namespace = %q, want cluster-wide", got.Namespace)
+	}
+	want := provider.LabelManaged + "=" + provider.LabelManagedYes
+	if got.LabelSelector == nil || got.LabelSelector.String() != want {
+		t.Errorf("LabelSelector = %v, want %s", got.LabelSelector, want)
+	}
+}
+
+func TestCheck_failureReasons(t *testing.T) {
+	t.Parallel()
+	pods := schema.GroupResource{Resource: "pods"}
+	ended, cancel := context.WithCancel(context.Background())
+	cancel()
+
+	tests := []struct {
+		name string
+		ctx  context.Context
+		err  error
+		want string
+	}{
+		{"forbidden", context.Background(), apierrors.NewForbidden(pods, "", errors.New("rbac denied")), provider.ReasonForbidden},
+		{"unauthorized", context.Background(), apierrors.NewUnauthorized("token expired"), provider.ReasonUnauthorized},
+		{"api timeout", context.Background(), apierrors.NewTimeoutError("slow etcd", 1), provider.ReasonTimeout},
+		{"server timeout", context.Background(), apierrors.NewServerTimeout(pods, "list", 1), provider.ReasonTimeout},
+		{"context ended", ended, context.Canceled, provider.ReasonTimeout},
+		{"internal error", context.Background(), apierrors.NewInternalError(errors.New("etcd down")), provider.ReasonAPIError},
+		{"connection refused", context.Background(), &url.Error{
+			Op: "Get", URL: "https://10.96.0.1:443/api/v1/pods", Err: syscall.ECONNREFUSED,
+		}, provider.ReasonUnreachable},
+	}
+	for _, tc := range tests {
+		t.Run(tc.name, func(t *testing.T) {
+			t.Parallel()
+			p := newCheckProvider(func(context.Context, client.WithWatch, client.ObjectList, ...client.ListOption) error {
+				return tc.err
+			})
+
+			got := p.Check(tc.ctx)
+
+			if len(got) != 1 {
+				t.Fatalf("Check() returned %d targets, want 1", len(got))
+			}
+			if got[0].Ready || got[0].Reason != tc.want || got[0].Message != tc.err.Error() {
+				t.Errorf("Check() = %+v, want not ready, reason %s, message %q", got[0], tc.want, tc.err.Error())
+			}
+		})
+	}
+}
@@ -36,6 +36,8 @@ const defaultImage = "ubuntu/squid:6.6-24.04_edge"
 type Provider struct {
 	client client.Client
 	image  string
+	// apiHost is the API server URL, shown by Check; empty in tests.
+	apiHost string
 }

 // New builds a kubernetes Provider from its config block. Satisfies
@@ -64,7 +66,9 @@ func NewWithTransportWrapper(_ context.Context, cfg provider.ProviderConfig, wra
 	if err != nil {
 		return nil, fmt.Errorf("kubernetes provider %q: building client: %w", cfg.Name, err)
 	}
-	return newWithClient(c, cfg), nil
+	p := newWithClient(c, cfg)
+	p.apiHost = restCfg.Host
+	return p, nil
 }

 // newWithClient builds a Provider around an already-constructed client,