diff --git a/pkg/agent/resourcemanager/fetcher/kubelet/topology/topology_adapter.go b/pkg/agent/resourcemanager/fetcher/kubelet/topology/topology_adapter.go
index 00548593c..ddd989353 100644
--- a/pkg/agent/resourcemanager/fetcher/kubelet/topology/topology_adapter.go
+++ b/pkg/agent/resourcemanager/fetcher/kubelet/topology/topology_adapter.go
@@ -53,6 +53,7 @@ import (
 
 const (
 	podResourcesClientTimeout    = 10 * time.Second
+	getTopologyZonesTimeout      = 10 * time.Second
 	podResourcesClientMaxMsgSize = 1024 * 1024 * 16
 )
 
@@ -139,6 +140,8 @@ func (p *topologyAdapterImpl) GetTopologyZones(parentCtx context.Context) ([]*no
 	// always force getting pod list instead of cache
 	ctx := context.WithValue(parentCtx, metaserverpod.BypassCacheKey, metaserverpod.BypassCacheTrue)
 
+	ctx, cancel := context.WithTimeout(ctx, getTopologyZonesTimeout)
+	defer cancel()
 	podList, err := p.metaServer.GetPodList(ctx, nil)
 	if err != nil {
 		return nil, errors.Wrap(err, "get pod list from metaServer failed")
diff --git a/pkg/agent/resourcemanager/fetcher/plugin/endpoint.go b/pkg/agent/resourcemanager/fetcher/plugin/endpoint.go
index a4b4c0b11..eb6e1c399 100644
--- a/pkg/agent/resourcemanager/fetcher/plugin/endpoint.go
+++ b/pkg/agent/resourcemanager/fetcher/plugin/endpoint.go
@@ -33,6 +33,7 @@ import (
 
 const (
 	dialRemoteEndpointTimeout = 10 * time.Second
+	getReportContentTimeout   = 10 * time.Second
 )
 
 // ListAndWatchCallback should be called when plugins report info update.
@@ -173,7 +174,9 @@ func (e *remoteEndpointImpl) GetReportContent(c context.Context) (*v1alpha1.GetR
 		return nil, fmt.Errorf("endpoint %v has been stopped", e.pluginName)
 	}
 
-	resp, err := e.client.GetReportContent(c, &v1alpha1.Empty{})
+	ctx, cancel := context.WithTimeout(c, getReportContentTimeout)
+	defer cancel()
+	resp, err := e.client.GetReportContent(ctx, &v1alpha1.Empty{})
 	if err == nil {
 		e.setCache(resp)
 	}
diff --git a/pkg/agent/sysadvisor/plugin/overcommitmentaware/reporter/reporter.go b/pkg/agent/sysadvisor/plugin/overcommitmentaware/reporter/reporter.go
index 05e921c4e..292a9ee73 100644
--- a/pkg/agent/sysadvisor/plugin/overcommitmentaware/reporter/reporter.go
+++ b/pkg/agent/sysadvisor/plugin/overcommitmentaware/reporter/reporter.go
@@ -153,7 +153,7 @@ func (o *OvercommitRatioReporterPlugin) Stop() error {
 // GetReportContent get overcommitment ratio from manager directly.
 // Since the metrics collected by Manager are already an average within a time period,
 // we expect a faster response to node load fluctuations to avoid excessive overcommit of online resources.
-func (o *OvercommitRatioReporterPlugin) GetReportContent(_ context.Context, _ *v1alpha1.Empty) (*v1alpha1.GetReportContentResponse, error) {
+func (o *OvercommitRatioReporterPlugin) GetReportContent(ctx context.Context, _ *v1alpha1.Empty) (*v1alpha1.GetReportContentResponse, error) {
 	response := &v1alpha1.GetReportContentResponse{
 		Content: []*v1alpha1.ReportContent{},
 	}
@@ -173,7 +173,7 @@ func (o *OvercommitRatioReporterPlugin) GetReportContent(_ context.Context, _ *v
 	response.Content = append(response.Content, overcommitRatioContent)
 
 	// get topologyProvider and guaranteed cpus
-	topologyProviderContent, err := o.getTopologyProviderReportContent(o.ctx)
+	topologyProviderContent, err := o.getTopologyProviderReportContent(ctx)
 	if err != nil {
 		return nil, err
 	}