diff --git a/deploy/gcp/main.tf b/deploy/gcp/main.tf
index f833830d0ed..aa906058d62 100644
--- a/deploy/gcp/main.tf
+++ b/deploy/gcp/main.tf
@@ -170,7 +170,7 @@ resource "google_container_node_pool" "monitor_pool" {
   cluster            = "${google_container_cluster.cluster.name}"
   location           = "${google_container_cluster.cluster.location}"
   name               = "monitor-pool"
-  initial_node_count = "1"
+  initial_node_count = "${var.monitor_count}"
 
   node_config {
     machine_type = "${var.monitor_instance_type}"
@@ -251,12 +251,15 @@ resource "null_resource" "get-credentials" {
       KUBECONFIG = "${local.kubeconfig}"
     }
   }
+
   provisioner "local-exec" {
     when = "destroy"
+
     command = <<EOS
 kubectl get pvc -n tidb -o jsonpath='{.items[*].spec.volumeName}'|fmt -1 | xargs -I {} kubectl patch pv {} -p '{"spec":{"persistentVolumeReclaimPolicy":"Delete"}}'
 kubectl delete namespace tidb
 EOS
+
     environment {
       KUBECONFIG = "${local.kubeconfig}"
     }
diff --git a/deploy/gcp/variables.tf b/deploy/gcp/variables.tf
index c4eec49012e..99097fdac9c 100644
--- a/deploy/gcp/variables.tf
+++ b/deploy/gcp/variables.tf
@@ -35,6 +35,11 @@ variable "tidb_count" {
   default     = 1
 }
 
+variable "monitor_count" {
+  description = "Number of monitor nodes per availability zone"
+  default     = 1
+}
+
 variable "pd_instance_type" {
   default = "n1-standard-4"
 }