diff --git a/cmd/validator.go b/cmd/validator.go
index a914c13..874c9ab 100644
--- a/cmd/validator.go
+++ b/cmd/validator.go
@@ -316,10 +316,38 @@ var validatorAddAutoRenewedCmd = &cobra.Command{
 		}
 
 		fmt.Printf("TX ID: %s\n", txID)
+		warnIfValidatorNotConnected(ctx, netConfig.RPCURL, nodeID)
 		return nil
 	},
 }
 
+// warnIfValidatorNotConnected reports when the network cannot see the node that
+// was just staked. Auto-renewal is conditioned on reward eligibility, so a node
+// the network never reaches is removed at the first cycle boundary and forfeits
+// that cycle's rewards instead of renewing. The transaction has already been
+// accepted by this point, so every failure here is advisory only.
+func warnIfValidatorNotConnected(ctx context.Context, rpcURL string, nodeID ids.NodeID) {
+	health, err := pchain.GetValidatorHealth(ctx, rpcURL, nodeID)
+	if err != nil {
+		fmt.Printf("\nNote: could not check whether %s is connected: %v\n", nodeID, err)
+		return
+	}
+	if !health.Found {
+		fmt.Printf("\nNote: %s is not in the validator set yet. Confirm the node is connected once it appears.\n", nodeID)
+		return
+	}
+	if health.Connected {
+		return
+	}
+
+	fmt.Printf("\nWarning: the network does not currently see %s as connected.\n", nodeID)
+	fmt.Println("Auto-renewal requires meeting the uptime requirement at each cycle boundary. A node")
+	fmt.Println("that stays unreachable is removed at the end of the cycle, forfeiting that cycle's")
+	fmt.Println("rewards, rather than renewing.")
+	fmt.Println("Check that the node is running with the BLS key used here, that --public-ip is set,")
+	fmt.Println("and that inbound TCP port 9651 is reachable.")
+}
+
 var validatorSetAutoConfigCmd = &cobra.Command{
 	Use:   "set-auto-renewed-config",
 	Short: "Set auto-renewed validator config (SetAutoRenewedValidatorConfigTx)",
diff --git a/pkg/pchain/validator_health.go b/pkg/pchain/validator_health.go
new file mode 100644
index 0000000..11d53c8
--- /dev/null
+++ b/pkg/pchain/validator_health.go
@@ -0,0 +1,72 @@
+package pchain
+
+import (
+	"context"
+	"fmt"
+	"strconv"
+
+	"github.com/ava-labs/avalanchego/ids"
+	"github.com/ava-labs/avalanchego/utils/constants"
+	"github.com/ava-labs/avalanchego/vms/platformvm"
+)
+
+// ValidatorHealth is how the network currently sees a primary network validator.
+//
+// It matters most for auto-renewed validators: renewal is conditioned on reward
+// eligibility, so a validator the network cannot reach is removed at the next
+// cycle boundary rather than renewed, and forfeits that cycle's rewards.
+type ValidatorHealth struct {
+	// Found is false when nodeID is not in the current validator set.
+	Found     bool
+	Connected bool
+	// UptimePercent is the queried node's view of this validator's uptime. It is
+	// 0 for a validator that has only just started validating, so it is only
+	// meaningful once the validator has been in the set for a while.
+	UptimePercent float64
+}
+
+// GetValidatorHealth reports the connectivity and uptime that
+// platform.getCurrentValidators returns for nodeID on the primary network.
+func GetValidatorHealth(ctx context.Context, rpcURL string, nodeID ids.NodeID) (*ValidatorHealth, error) {
+	client := platformvm.NewClient(rpcURL)
+	args := &platformvm.GetCurrentValidatorsArgs{
+		SubnetID: constants.PrimaryNetworkID,
+		NodeIDs:  []ids.NodeID{nodeID},
+	}
+	reply := &getCurrentValidatorsHealthReply{}
+	if err := client.Requester.SendRequest(ctx, "platform.getCurrentValidators", args, reply); err != nil {
+		return nil, fmt.Errorf("failed to fetch current validators: %w", err)
+	}
+
+	for _, validator := range reply.Validators {
+		if validator.NodeID != nodeID.String() {
+			continue
+		}
+		health := &ValidatorHealth{
+			Found:     true,
+			Connected: validator.Connected,
+		}
+		// Uptime is omitted for subnet validators and may be absent on nodes that
+		// do not track it; treat a missing value as 0 rather than an error.
+		if validator.Uptime != "" {
+			uptime, err := strconv.ParseFloat(validator.Uptime, 64)
+			if err != nil {
+				return nil, fmt.Errorf("invalid uptime %q for %s: %w", validator.Uptime, nodeID, err)
+			}
+			health.UptimePercent = uptime
+		}
+		return health, nil
+	}
+
+	return &ValidatorHealth{}, nil
+}
+
+type getCurrentValidatorsHealthReply struct {
+	Validators []validatorHealthEntry `json:"validators"`
+}
+
+type validatorHealthEntry struct {
+	NodeID    string `json:"nodeID"`
+	Connected bool   `json:"connected"`
+	Uptime    string `json:"uptime"`
+}
diff --git a/pkg/pchain/validator_health_test.go b/pkg/pchain/validator_health_test.go
new file mode 100644
index 0000000..a5e616a
--- /dev/null
+++ b/pkg/pchain/validator_health_test.go
@@ -0,0 +1,106 @@
+package pchain
+
+import (
+	"context"
+	"strings"
+	"testing"
+
+	"github.com/ava-labs/avalanchego/ids"
+)
+
+func TestGetValidatorHealth(t *testing.T) {
+	nodeID := ids.GenerateTestNodeID()
+	otherNodeID := ids.GenerateTestNodeID()
+
+	tests := []struct {
+		name          string
+		validators    []map[string]any
+		wantFound     bool
+		wantConnected bool
+		wantUptime    float64
+	}{
+		{
+			name: "connected validator",
+			validators: []map[string]any{{
+				"nodeID":    nodeID.String(),
+				"connected": true,
+				"uptime":    "99.5000",
+			}},
+			wantFound:     true,
+			wantConnected: true,
+			wantUptime:    99.5,
+		},
+		{
+			name: "disconnected validator",
+			validators: []map[string]any{{
+				"nodeID":    nodeID.String(),
+				"connected": false,
+				"uptime":    "0.0000",
+			}},
+			wantFound:     true,
+			wantConnected: false,
+		},
+		{
+			name: "missing uptime is treated as zero",
+			validators: []map[string]any{{
+				"nodeID":    nodeID.String(),
+				"connected": true,
+			}},
+			wantFound:     true,
+			wantConnected: true,
+		},
+		{
+			name:       "not in the validator set",
+			validators: []map[string]any{},
+		},
+		{
+			name: "other node ids are ignored",
+			validators: []map[string]any{{
+				"nodeID":    otherNodeID.String(),
+				"connected": true,
+				"uptime":    "100.0000",
+			}},
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			var gotParams string
+			server := newCurrentValidatorsServer(t, &gotParams, tt.validators)
+			defer server.Close()
+
+			health, err := GetValidatorHealth(context.Background(), server.URL, nodeID)
+			if err != nil {
+				t.Fatalf("GetValidatorHealth() error = %v", err)
+			}
+			if health.Found != tt.wantFound {
+				t.Errorf("Found = %v, want %v", health.Found, tt.wantFound)
+			}
+			if health.Connected != tt.wantConnected {
+				t.Errorf("Connected = %v, want %v", health.Connected, tt.wantConnected)
+			}
+			if health.UptimePercent != tt.wantUptime {
+				t.Errorf("UptimePercent = %v, want %v", health.UptimePercent, tt.wantUptime)
+			}
+			// The query must be scoped to the node and the primary network, so the
+			// reply cannot be diluted by the whole validator set.
+			if !strings.Contains(gotParams, nodeID.String()) {
+				t.Errorf("request params = %s, want it to contain %s", gotParams, nodeID)
+			}
+		})
+	}
+}
+
+func TestGetValidatorHealthInvalidUptime(t *testing.T) {
+	nodeID := ids.GenerateTestNodeID()
+	server := newCurrentValidatorsServer(t, nil, []map[string]any{{
+		"nodeID":    nodeID.String(),
+		"connected": true,
+		"uptime":    "not-a-number",
+	}})
+	defer server.Close()
+
+	if _, err := GetValidatorHealth(context.Background(), server.URL, nodeID); err == nil {
+		t.Fatal("GetValidatorHealth() error = nil, want an error for an unparseable uptime")
+	}
+}