cli: handle standalone Envoy startup failures (#1994)

**Description**

In standalone mode, handle EG infrastructure errors and exit gracefully
instead of hanging.

**Related Issues/PRs (if applicable)**

Related to: #1304

**Special notes for reviewers (if applicable)**

N/A

---------

Signed-off-by: Ignasi Barrera <nacx@apache.org>
Co-authored-by: Aaron Choo <achoo30@bloomberg.net>

Author: nacx

cmd/aigw/run.go

@@ -19,6 +19,7 @@ import (
 	"strings"
 	"time"
 
+	egv1a1 "github.com/envoyproxy/gateway/api/v1alpha1"
 	"github.com/envoyproxy/gateway/cmd/envoy-gateway/root"
 	egextension "github.com/envoyproxy/gateway/proto/extension"
 	"github.com/go-logr/logr"
@@ -206,8 +207,9 @@ func run(ctx context.Context, c *cmdRun, o *runOpts, stdout, stderr io.Writer) e
 	//
 	// Now running the `envoy-gateway` CLI alternative below by passing `--config-path` to `egConfigPath`.
 	// Then the agent will read the resources from the file pointed inside the config and start the Envoy process.
+	runnerErrorHandler := newRunnerErrorHandler(os.Stderr, serverCancel)
+	server := root.GetRootCommand(runnerErrorHandler)
 
-	server := root.GetRootCommand(nil)
 	// TODO: enable the log by default after the issue is resolved: https://github.com/envoyproxy/gateway/issues/6596
 	if c.Debug {
 		server.SetOut(stdout)
@@ -491,6 +493,18 @@ func (runCtx *runCmdContext) tryFindEnvoyListenerPort(gw *gwapiv1.Gateway) int {
 	return int(gw.Spec.Listeners[0].Port)
 }
 
+// newRunnerErrorHandler returns a callback that triggers a graceful shutdown
+// (via cancelFunc) when a critical Envoy Gateway runner fails.
+func newRunnerErrorHandler(stderr io.Writer, cancelFunc context.CancelFunc) func(string, error) {
+	return func(runner string, err error) {
+		if runner == string(egv1a1.LogComponentProviderRunner) ||
+			runner == string(egv1a1.LogComponentInfrastructureRunner) {
+			_, _ = fmt.Fprintf(stderr, "exiting on %s runner error: %v\n", runner, err)
+			cancelFunc()
+		}
+	}
+}
+
 func maybeResolveHome(p string) string {
 	if strings.HasPrefix(p, "~/") {
 		home, err := os.UserHomeDir()

cmd/aigw/run_test.go

@@ -288,6 +288,68 @@ func Test_newEnvoyMiddleware(t *testing.T) {
 	}
 }
 
+func Test_newRunnerErrorHandler(t *testing.T) {
+	tests := []struct {
+		name           string
+		runner         string
+		expectCancel   bool
+		expectInStderr bool
+	}{
+		{
+			name:           "provider runner triggers graceful shutdown",
+			runner:         "provider",
+			expectCancel:   true,
+			expectInStderr: true,
+		},
+		{
+			name:           "infrastructure runner triggers graceful shutdown",
+			runner:         "infrastructure",
+			expectCancel:   true,
+			expectInStderr: true,
+		},
+		{
+			name:         "gateway-api runner does not trigger shutdown",
+			runner:       "gateway-api",
+			expectCancel: false,
+		},
+		{
+			name:         "xds-translator runner does not trigger shutdown",
+			runner:       "xds-translator",
+			expectCancel: false,
+		},
+		{
+			name:         "unknown runner does not trigger shutdown",
+			runner:       "unknown",
+			expectCancel: false,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			ctx, cancel := context.WithCancel(t.Context())
+			defer cancel()
+
+			var stderr bytes.Buffer
+			handler := newRunnerErrorHandler(&stderr, cancel)
+
+			mockErr := errors.New("something went wrong")
+			handler(tt.runner, mockErr)
+
+			if tt.expectCancel {
+				require.Error(t, ctx.Err(), "context should be cancelled")
+			} else {
+				require.NoError(t, ctx.Err(), "context should not be cancelled")
+			}
+			if tt.expectInStderr {
+				require.Contains(t, stderr.String(), tt.runner)
+				require.Contains(t, stderr.String(), "something went wrong")
+			} else {
+				require.Empty(t, stderr.String())
+			}
+		})
+	}
+}
+
 func findFlagValue(args []string, flag string) string {
 	for i, arg := range args {
 		if arg == flag && i+1 < len(args) {