Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Manual backport of docs: Consul DNS views on Kubernetes (#21802) Beta into release/1.20.x #21835

Merged
merged 1 commit into from
Oct 16, 2024
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Original file line number Diff line number Diff line change
Expand Up @@ -5,16 +5,15 @@ description: >-
Use a k8s ConfigMap to configure KubeDNS or CoreDNS so that you can use Consul's `<service-name>.service.consul` syntax for queries and other DNS requests. In Kubernetes, this process uses either stub-domain or proxy configuration.
---

# Resolve Consul DNS Requests in Kubernetes
# Resolve Consul DNS requests in Kubernetes

One of the primary query interfaces to Consul is the
[DNS interface](/consul/docs/services/discovery/dns-overview). You can configure Consul DNS in
This topic describes how to configure Consul DNS in
Kubernetes using a
[stub-domain configuration](https://kubernetes.io/docs/tasks/administer-cluster/dns-custom-nameservers/#configure-stub-domain-and-upstream-dns-servers)
if using KubeDNS or a [proxy configuration](https://coredns.io/plugins/forward/) if using CoreDNS.

Once configured, DNS requests in the form `<consul-service-name>.service.consul` will
resolve for services in Consul. This will work from all Kubernetes namespaces.
resolve for services in Consul. This works from all Kubernetes namespaces.

-> **Note:** If you want requests to just `<consul-service-name>` (without the `.service.consul`) to resolve, then you'll need
to turn on [Consul to Kubernetes Service Sync](/consul/docs/k8s/service-sync#consul-to-kubernetes).
Expand Down
102 changes: 102 additions & 0 deletions website/content/docs/k8s/dns/views/enable.mdx
Original file line number Diff line number Diff line change
@@ -0,0 +1,102 @@
---
layout: docs
page_title: Enable Consul DNS proxy for Kubernetes
description: ->
Learn how to schedule a Consul DNS proxy for a Kubernetes Pod so that your services can return Consul DNS results for service discovery.
---

# Enable Consul DNS proxy for Kubernetes

This page describes the process to deploy a Consul DNS proxy in a Kubernetes Pod so that Services can resolve Consul DNS requests. For more information, refer to [Consul DNS views for Kubernetes](/consul/docs/k8s/dns/views).

## Prerequisites

You must meet the following minimum application versions to enable the Consul DNS proxy for Kubernetes:

- Consul v1.20.0 or higher
- Either Consul on Kubernetes or the Consul Helm chart, v1.6.0 or higher

## Update Helm values

To enable the Consul DNS proxy, add the required [Helm values](/consul/docs/k8s/helm) to your Consul on Kubernetes deployment.

```yaml
connectInject:
enabled: true
dns:
enabled: true
proxy: true
```

### ACLs

We recommend you create a dedicated [ACL token with DNS permissions](/consul/docs/security/acl/tokens/create/create-a-dns-token) for the Consul DNS proxy. The Consul DNS proxy requires these ACL permissions.

```hcl
node_prefix "" {
policy = "read"
}

service_prefix "" {
policy = "read"
}
```

You can manage ACL tokens with Consul on Kubernetes, or you can configure the DNS proxy to access a token stored in Kubernetes secret. To use a Kubernetes secret, add the following configuration to your Helm chart.

```yaml
dns:
proxy:
aclToken:
secretName: <Consul-DNS-Token>
secretKey: <Token-Value>
```

## Retrieve Consul DNS proxy's address

To look up the IP address for the Consul DNS proxy in the Kubernetes Pod, run the following command.

```shell-session
$ kubectl get services –-all-namespaces --selector="app=consul,component=dns-proxy" --output jsonpath='{.spec.clusterIP}'
10.96.148.46
```

Use this address when you update the ConfigMap resource.

## Update Kubernetes ConfigMap

Create or update a [ConfigMap object in the Kubernetes cluster](https://kubernetes.io/docs/concepts/configuration/configmap/) so that Kubernetes forwards DNS requests with the `.consul` domain to the IP address of the Consul DNS proxy.

The following example of a `coredns-custom` ConfigMap configures Kubernetes to forward Consul DNS requests in the cluster to the Consul DNS Proxy running on `10.96.148.46`. This resource modifies the CoreDNS without modifications to the original `Corefile`.

```yaml
kind: ConfigMap
metadata:
name: coredns-custom
namespace: kube-system
data:
consul.server: |
consul:53 {
errors
cache 30
forward . 10.96.148.46
reload
}
```

After updating the DNS configuration, perform a rolling restart of the CoreDNS.

```shell-session
kubectl -n kube-system rollout restart deployment coredns
```

For more information about using a `coredns-custom` resource, refer to the [Rewrite DNS guide in the Azure documentation](https://learn.microsoft.com/en-us/azure/aks/coredns-custom#rewrite-dns). For general information about modifying a ConfigMap, refer to [the Kubernetes documentation](https://kubernetes.io/docs/tasks/administer-cluster/dns-custom-nameservers/#coredns).

## Next steps

After you enable the Consul DNS proxy, services in the Kubernetes cluster can resolve Consul DNS addresses.

- To learn more about Consul DNS for service discovery, refer to [DNS usage overview](/consul/docs/services/discovery/dns-overview).
- If your datacenter has ACLs enabled, create a [Consul ACL token](/consul/docs/security/acl/tokens) for the Consul DNS proxy and then restart the DNS proxy.
- To enable service discovery across admin partitions, [export services between partitions](/consul/docs/connect/config-entries/exported-services).
- To use Consul DNS for service discovery with other runtimes, across cloud regions, or between cloud providers, [establish a cluster peering connection](/consul/docs/k8s/connect/cluster-peering/usage/establish-peering).
48 changes: 48 additions & 0 deletions website/content/docs/k8s/dns/views/index.mdx
Original file line number Diff line number Diff line change
@@ -0,0 +1,48 @@
---
layout: docs
page_title: Consul DNS views for Kubernetes
description: ->
Kubernetes clusters can use the Consul DNS proxy to return service discovery results from the Consul catalog. Learn about how to configure your k8s cluster so that applications can resolve Consul DNS addresses without gossip communication.
---

# Consul DNS views for Kubernetes

This topic describes how to schedule a dedicated Consul DNS proxy in a Kubernetes Pod so that applications in Kubernetes can resolve Consul DNS addresses. You can use the Consul DNS proxy to enable service discovery across admin partitions in Kubernetes deployments without needing to deploy Consul client agents.

## Introduction

Kubernetes operators typically choose networking tools such as [kube-dns](https://kubernetes.io/docs/concepts/services-networking/dns-pod-service/) or [CoreDNS](https://kubernetes.io/docs/tasks/administer-cluster/coredns/) for their service discovery operations, and choose to bypass Consul DNS entirely. These DNS options are often sufficient for service networking operations within a single Kubernetes cluster.

Consul on Kubernetes supports [configuring Kubernetes to resolve Consul DNS](/consul/docs/k8s/dns). However, two common challenges result when you rely on these configurations:

- Kubernetes requires Consul to use gossip communication with agents or dataplanes in order to enable Consul DNS.
- Consul requires that admin partitions be included in the DNS address. Otherwise, DNS queries assume the `default` partition by default.

The `consul-dns` proxy does not require the presence of Consul client agents or Consul dataplanes, removing gossip communication as a requirement for Consul DNS on Kubernetes. The proxy is also designed for deployment in a Kubernetes cluster with [external servers enabled](/consul/docs/k8s/deployment-configurations/servers-outside-kubernetes). When a cluster runs in a non-default admin partition and uses the proxy to query external servers, Consul automatically recognizes the admin partition that originated the request and returns service discovery results scoped to that specific admin partition.

To use Consul DNS for service discovery on Kubernetes, deploy a `dns-proxy` service in each Kubernetes Pod that needs to resolve Consul DNS. Kubernetes sends all DNS requests to the Kubernetes controller first. The controller forwards requests for the `.consul` domain to the `dns-proxy` service, which then queries the Consul catalog and returns service discovery results.

## Workflows

The process to enable Consul DNS views for service discovery in Kubernetes deployments consists of the following steps:

1. In a cluster configured to use [external Consul servers](/consul/docs/k8s/deployment-configurations/servers-outside-kubernetes), update the Helm values for your Consul on Kubernetes deployment so that `dns.proxy.enabled=true`. When you apply the updated configuration, Kubernetes deploys the Consul DNS proxy.
1. Look up the IP address for the Consul DNS proxy in the Kubernetes cluster.
1. Update the ConfigMap resource in the Kubernetes cluster so that it forwards requests for the `.consul` domain to the IP address of the Consul DNS proxy.

For more information about the underlying concepts described in this workflow, refer to [DNS forwarding overview](/consul/docs/services/discovery/dns-forwarding).

## Benefits

Consul on Kubernetes currently uses [Consul dataplanes](/consul/docs/connect/dataplane) by default. These lightweight processes provide Consul access to the sidecar proxies in the service mesh, but leave Kubernetes in charge of most other service discovery and service mesh operations.

- **Use Kubernetes DNS and Consul DNS in a single deployment**. The Consul DNS proxy enables any application in a Pod to resolve an address through Consul DNS without disrupting the underlying Kubernetes DNS functionality.
- **Consul service discovery using fewer resources**. When you use the Consul DNS proxy for service discovery, you do not need to schedule Consul client agents or dataplanes as sidecars. One Kubernetes Service that uses the same resources as a single Consul dataplane provides Pods access to the Consul service catalog.
- **Consul DNS without gossip communication**. The Consul DNS service runs on both Consul server and Consul client agents, which use [gossip communication](/consul/docs/security/encryption/gossip) to ensure that service discovery results are up-to-date. The Consul DNS proxy provides access to Consul DNS without the security overhead of agent-to-agent gossip.

## Constraints and limitations

If you experience issues using the Consul DNS proxy for Kubernetes, refer to the following list of technical constraints and limitations.

- You must use Kubernetes as your runtime to use the Consul DNS proxy. You cannot schedule the Consul DNS proxy in other container-based environments.
- To perform DNS lookups on other admin partitions, you must [export services between partitions](/consul/docs/connect/config-entries/exported-services) before you can query them.
27 changes: 19 additions & 8 deletions website/content/docs/services/discovery/dns-cache.mdx
Original file line number Diff line number Diff line change
@@ -1,22 +1,33 @@
---
layout: docs
page_title: Enable dynamic DNS queries
page_title: Scale Consul DNS
description: ->
You tune Consul DNS query handling to balance between current information and reducing request response time. Learn how to enable caching by modifying TTL values, how to return stale results from the DNS cache, and how to configure Consul for negative response caching.
---

# DNS caching
# Scale Consul DNS

This page describes the process to return cached results in response to DNS lookups. Consul agents can use DNS caching to reduce response time, but might provide stale information in the process.

## Introduction
## Scaling techniques

By default, Consul serves all DNS results with a `0` TTL value, which prevents any
caching. This configuration returns the most recent information because each DNS lookup
runs every time. However, this configuration adds latency to each lookup and can potentially
exhaust the query throughput of a datacenter.
By default, Consul serves all DNS results with a `0` TTL value so that it returns the most recent information. When operating at scale, this configuration may result in additional latency because servers must respond to every DNS query. There are several strategies for distributing this burden in your datacenter:

There are several ways you can modify to fine-tune Consul DNS lookup behavior to best suit your network's requirements.
- [Allow Stale Reads](#stale-reads). Allows other servers besides the leader to answer the query rather than forwarding it to the leader.
- [Configure DNS TTLs](#ttl-values). Configure DNS time-to-live (TTL) values for nodes or services so that the DNS subsystem on the container’s operating system can cache responses. Services then resolve DNS queries locally without any external requests.
- [Add Read Replicas](/consul/docs/enterprise/read-scale). Enterprise users can use read replicas, which are additional Consul servers that replicate cluster data without participating in the Raft quorum.
- [Use Cache to prevent server requests](/consul/docs/agent/config/config-files#dns_use_cache). Configure the Consul client to use its agent cache to subscribe to events about a service or node. After you establish the watch, the local Consul client agent can resolve DNS queries about the service or node without querying Consul servers.

The following table describes the availability of each scaling technique depending on whether you configure Consul to offload DNS requests from the cluster leader to a client agent, dataplane, or DNS proxy.

| Scaling technique | Supported by client agents | Supported by dataplanes | Supported by Consul DNS Proxy |
| :---------------------------------- | :---------------------------: | :---------------------------: | :---------------------------: |
| Configure DNS TTLs | &#9989; | &#9989; | &#9989; |
| Allow Stale Reads | &#9989; | &#9989; | &#9989; |
| Add Read Replicas | &#9989; | &#9989; | &#9989; |
| Use Cache to prevent server request | &#9989; | &#10060; | &#10060; |

For more information about considerations for Consul deployments that operate at scale, refer to [Operating Consul at Scale](/consul/docs/architecture/scale).

## TTL values ((#ttl))

Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -20,12 +20,13 @@ When configured with default values, Consul exposes the DNS interface on port `8
Instead of running Consul with an administrative or root account, you can forward appropriate queries to Consul, running on an unprivileged port, from another DNS server or using port redirect.

There are two configurations for a node's DNS forwarding behavior:

- **Conditional DNS forwarding**: the local DNS servers are configured to forward to Consul only queries relative to the `.consul` zone. All other queries are still served via the default DNS server in the node.
- **Full DNS forwarding**: Consul serves all DNS queries and forwards to a remote DNS server the ones outside `.consul` domain.

### Conditional DNS forwarding

We recommend the conditional DNS forwarding approach. This configuration lowers the Consul agent's resource consumption by limiting the number of DNS requests it handles.
We recommend the conditional DNS forwarding approach. This configuration lowers the Consul agent's resource consumption by limiting the number of DNS requests it handles.

![Consul DNS conditional forwarding - Only .consul requests are routed to Consul](/img/consul-dns-conditional-forwarding.png#light-theme-only)
![Consul DNS conditional forwarding - Only .consul requests are routed to Consul](/img/consul-dns-conditional-forwarding-dark.png#dark-theme-only)
Expand All @@ -43,7 +44,7 @@ This approach can be useful in scenarios where the Consul agent's node is alloca

This behavior is not enabled by default. Consul standard configuration only resolves DNS records inside the `.consul` zone. To enable DNS forwarding, you need to set the [recursors](/consul/docs/agent/config/config-files#recursors) option in your Consul configuration.

In this scenario, if a Consul DNS reply includes a `CNAME` record pointing outside the `.consul` top level domain, then the DNS reply only includes `CNAME` records by default.
In this scenario, if a Consul DNS reply includes a `CNAME` record pointing outside the `.consul` top level domain, then the DNS reply only includes `CNAME` records by default.

When `recursors` is set and the upstream resolver is functioning correctly, Consul tries to resolve CNAMEs and include any records (for example, A, AAAA, PTR) for them in its DNS reply. In these scenarios, Consul is used for full DNS forwarding and is able to serve queries for all domains.

Expand Down
Loading
Loading