Docs

Volcengine Kubernetes Engine

Workload FAQs

How do I update the CoreDNS replica count and resource quotas?

Issue description

In a VKE cluster, CoreDNS is installed by default as the domain resolution service of the cluster. When the cluster has a large number of nodes or experiences high DNS query pressure, you need to adjust the CoreDNS replica count, CPU/memory limits, or CPU/memory requests. This topic describes how to perform these operations.

Cause analysis

Scenario	Reason for adjustment
Adjust the number of replicas	The cluster has a large number of nodes. A single CoreDNS instance cannot handle the DNS query load, resulting in a performance bottleneck.
Adjust resource limits	The default resource configuration cannot meet the current cluster DNS query load, resulting in an OOM or CPU resource shortage error.

Solution

In VKE, CoreDNS is deployed as a system component in the kube-system namespace as a ‌Deployment‌. You can directly modify its configuration to adjust the ‌replica count‌ and ‌resource limits‌.

Method 1: Using the console (recommended)

In the left navigation pane, click Workload > Deployment.

Select the kube-system namespace and search for coredns. In the Actions column on the right, click ...> Update.

On the Update Deployment page, modify the following parameters. For more information, see Workload configuration.

Basic information

Number of instances: Adjust the number of instances to the desired value.

Container configuration

ResourceQuota: Modify the CPU and memory requests and limits as needed.

After you confirm the configuration, click Confirm to save the changes.

After the changes are saved, the system will automatically perform a rolling update of the corresponding CoreDNS Pod. You can view the Pod status on the Instances tab on the Deployment details page. The Running state indicates that the configuration changes have taken effect.

Method 2: Using kubectl

If you prefer using the command line, follow these steps:

Connect to the cluster via kubectl. For details, see Connecting to a cluster.

View the current CoreDNS configuration in the cluster.

# Check the current CoreDNS replica count
kubectl get deployment -n kube-system coredns

# Check the current resource configuration
kubectl get deployment -n kube-system coredns -o yaml | grep -A 10 resources

Modify the CoreDNS configuration.

Modify the CoreDNS replica count:

# Change the replica count to 2 (adjust it as needed in your actual scenarios)
kubectl scale deployment coredns -n kube-system --replicas=2

Modify CoreDNS requests or limits:

Use the edit command (recommended)
Use the patch command
Run the following command to edit the CoreDNS configuration.

kubectl edit deployment coredns -n kube-system

In the editor, find the spec.template.spec.containers[0].resources section and change the value of limits.

resources:
requests:
cpu: 100m
memory: 128Mi
limits:
cpu: 250m # Change to the desired value.
memory: 256Mi # Change to the desired value.
# Example: Change CPU limit to 500m and memory limit to 512Mi
kubectl patch deployment coredns -n kube-system -p '
{
"spec": {
"template": {
"spec": {
"containers": [
{
"name": "coredns",
"resources": {
"requests": {
"cpu": "100m",
"memory": "128Mi"
},
"limits": {
"cpu": "500m",
"memory": "512Mi"
}
}
}
]
}
}
}
}'

Verify the changes.

# Check whether the replica count has been updated.
kubectl get pods -n kube-system -l k8s-app=kube-dns

# Check whether the resource configuration has been updated.
kubectl get pod -n kube-system -l k8s-app=kube-dns -o jsonpath='{.items[0].spec.containers[0].resources}' && echo

Recommended configuration

The recommended CoreDNS configuration varies with cluster node count, as described below.

warning

The following recommended settings are for reference only. You need to adjust your configuration based on the actual number of DNS queries. If the cluster has the node-local-dns component enabled, the pressure on CoreDNS will be greatly reduced, and the number of replicas can be appropriately decreased. For more details, see How do I check whether the node-local-dns component is being used and Using NodeLocal DNSCache as a DNS caching agent.

Cluster node count	Recommended replica count	Recommended limits
1 to 10 nodes	2 replicas	CPU: 250 millicores Memory: 256 MiB
10 to 50 nodes	4 replicas	CPU: 500 millicores Memory: 512 MiB
50 to 100 nodes	6 to 8 replicas	CPU: 1 core Memory: 1 GiB
100 nodes or more	8 replicas or more	CPU: 2 cores or more Memory: 2 GiB or more

Last updated: 2026.07.23 14:31:47