diff --git a/base/gfspvgmgr/virtual_group_manager.go b/base/gfspvgmgr/virtual_group_manager.go
index c0fa9d157..65151b919 100644
--- a/base/gfspvgmgr/virtual_group_manager.go
+++ b/base/gfspvgmgr/virtual_group_manager.go
@@ -19,6 +19,7 @@ import (
 	"github.com/bnb-chain/greenfield-storage-provider/core/consensus"
 	"github.com/bnb-chain/greenfield-storage-provider/core/vgmgr"
 	"github.com/bnb-chain/greenfield-storage-provider/pkg/log"
+	"github.com/bnb-chain/greenfield-storage-provider/pkg/metrics"
 	"github.com/bnb-chain/greenfield-storage-provider/util"
 	sptypes "github.com/bnb-chain/greenfield/x/sp/types"
 	virtualgrouptypes "github.com/bnb-chain/greenfield/x/virtualgroup/types"
@@ -33,9 +34,14 @@ const (
 	DefaultInitialGVGStakingStorageSize = uint64(2) * 1024 * 1024 * 1024 * 1024 // 2TB per GVG, chain side DefaultMaxStoreSizePerFamily is 64 TB
 	additionalGVGStakingStorageSize     = uint64(1) * 1024 * 1024 * 1024 * 1024 // 1TB
 
-	defaultSPCheckTimeout          = 3 * time.Second
-	defaultSPHealthCheckerInterval = 10 * time.Second
-	httpStatusPath                 = "/status"
+	defaultSPCheckTimeout               = 1 * time.Minute
+	defaultSPHealthCheckerInterval      = 10 * time.Second
+	defaultSPHealthCheckerRetryInterval = 1 * time.Second
+	defaultSPHealthCheckerMaxRetries    = 5
+	httpStatusPath                      = "/status"
+
+	SPHealthCheckerDuration = "check_sp_health_duration"
+	SPHealthCheckerFailure  = "check_sp_health_total_failure"
 
 	emptyGVGSafeDeletePeriod = int64(60) * 60 * 24
 )
@@ -785,30 +791,41 @@ func (checker *HealthChecker) checkSPHealth(sp *sptypes.StorageProvider) bool {
 		Transport: &http.Transport{
 			TLSClientConfig: &tls.Config{MinVersion: tls.VersionTLS12},
 		},
-		Timeout: defaultSPCheckTimeout * time.Second,
+		Timeout: defaultSPCheckTimeout,
 	}
 
 	// Create an HTTP request to test the validity of the endpoint
 	urlToCheck := fmt.Sprintf("%s%s", endpoint, httpStatusPath)
-	req, err := http.NewRequestWithContext(ctxTimeout, http.MethodGet, urlToCheck, nil)
-	if err != nil {
-		return false
-	}
+	for attempt := 0; attempt < defaultSPHealthCheckerMaxRetries; attempt++ {
+		start := time.Now()
+		req, err := http.NewRequestWithContext(ctxTimeout, http.MethodGet, urlToCheck, nil)
+		if err != nil {
+			log.CtxErrorw(context.Background(), "failed to create request", "sp", sp, "error", err)
+			return false
+		}
 
-	resp, err := client.Do(req)
-	if err != nil {
-		log.CtxErrorw(context.Background(), "failed to connect to sp", "sp", sp, "error", err)
-		return false
-	}
-	defer resp.Body.Close()
+		resp, err := client.Do(req)
+		duration := time.Since(start)
+		metrics.ReqTime.WithLabelValues(SPHealthCheckerDuration).Observe(duration.Seconds())
+		if err != nil {
+			log.CtxErrorw(context.Background(), "failed to connect to sp", "sp", sp, "error", err, "duration", duration)
+			time.Sleep(defaultSPHealthCheckerRetryInterval)
+			continue
+		}
+		defer resp.Body.Close()
 
-	if resp.StatusCode != http.StatusOK {
-		log.CtxErrorw(context.Background(), "failed to check sp healthy", "sp", sp, "http_status_code", resp.StatusCode, "resp_body", resp.Body)
-		return false
+		if resp.StatusCode == http.StatusOK {
+			log.CtxInfow(context.Background(), "succeed to check the sp healthy", "sp", sp, "duration", duration)
+			return true
+		} else {
+			metrics.ReqCounter.WithLabelValues(SPHealthCheckerFailure).Inc()
+			log.CtxErrorw(context.Background(), "failed to check sp healthy", "sp", sp, "http_status_code", resp.StatusCode, "resp_body", resp.Body, "duration", duration)
+			time.Sleep(defaultSPHealthCheckerRetryInterval)
+		}
 	}
 
-	log.CtxInfow(context.Background(), "succeed to check the sp healthy", "sp", sp)
-	return true
+	log.CtxErrorw(context.Background(), "failed to check sp healthy after retries", "sp", sp)
+	return false
 }
 
 func (checker *HealthChecker) Start() {