open-consul/website/content/docs/discovery/dns.mdx

359 lines
16 KiB
Plaintext

---
layout: docs
page_title: Find Services - DNS Interface
description: >-
One of the primary query interfaces for Consul is DNS. The DNS interface
allows applications to make use of service discovery without any high-touch
integration with Consul.
---
# DNS Interface
One of the primary query interfaces for Consul is DNS.
The DNS interface allows applications to make use of service
discovery without any high-touch integration with Consul.
For example, instead of making HTTP API requests to Consul,
a host can use the DNS server directly via name lookups
like `redis.service.us-east-1.consul`. This query automatically
translates to a lookup of nodes that provide the `redis` service,
are located in the `us-east-1` datacenter, and have no failing health checks.
It's that simple!
There are a number of configuration options that are important for the DNS interface,
specifically [`client_addr`](/docs/agent/options#client_addr),[`ports.dns`](/docs/agent/options#dns_port),
[`recursors`](/docs/agent/options#recursors),[`domain`](/docs/agent/options#domain),
[`alt_domain`](/docs/agent/options#alt_domain), and [`dns_config`](/docs/agent/options#dns_config).
By default, Consul will listen on 127.0.0.1:8600 for DNS queries in the `consul.`
domain, without support for further DNS recursion. Please consult the
[documentation on configuration options](/docs/agent/options),
specifically the configuration items linked above, for more details.
There are a few ways to use the DNS interface. One option is to use a custom
DNS resolver library and point it at Consul. Another option is to set Consul
as the DNS server for a node and provide a
[`recursors`](/docs/agent/options#recursors) configuration so that non-Consul queries
can also be resolved. The last method is to forward all queries for the "consul."
domain to a Consul agent from the existing DNS server. Review the
[DNS Forwarding tutorial](https://learn.hashicorp.com/tutorials/consul/dns-forwarding?utm_source=consul.io&utm_medium=docs) for examples.
You can experiment with Consul's DNS server on the command line using tools such as `dig`:
```shell-session
$ dig @127.0.0.1 -p 8600 redis.service.dc1.consul. ANY
```
-> **Note:** In DNS, all queries are case-insensitive. A lookup of `PostgreSQL.node.dc1.consul` will find all nodes named `postgresql`.
## Node Lookups
To resolve names, Consul relies on a very specific format for queries.
There are fundamentally two types of queries: node lookups and service lookups.
A node lookup, a simple query for the address of a named node, looks like this:
```text
<node>.node[.datacenter].<domain>
```
For example, if we have a `foo` node with default settings, we could
look for `foo.node.dc1.consul.` The datacenter is an optional part of
the FQDN: if not provided, it defaults to the datacenter of the agent.
If we know `foo` is running in the same datacenter as our local agent,
we can instead use `foo.node.consul.` This convention allows for terse
syntax where appropriate while supporting queries of nodes in remote
datacenters as necessary.
For a node lookup, the only records returned are A and AAAA records
containing the IP address, and TXT records containing the
`node_meta` values of the node.
```shell-session
$ dig @127.0.0.1 -p 8600 foo.node.consul ANY
; <<>> DiG 9.8.3-P1 <<>> @127.0.0.1 -p 8600 foo.node.consul ANY
; (1 server found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 24355
;; flags: qr aa rd; QUERY: 1, ANSWER: 1, AUTHORITY: 1, ADDITIONAL: 0
;; WARNING: recursion requested but not available
;; QUESTION SECTION:
;foo.node.consul. IN ANY
;; ANSWER SECTION:
foo.node.consul. 0 IN A 10.1.10.12
foo.node.consul. 0 IN TXT "meta_key=meta_value"
foo.node.consul. 0 IN TXT "value only"
;; AUTHORITY SECTION:
consul. 0 IN SOA ns.consul. postmaster.consul. 1392836399 3600 600 86400 0
```
By default the TXT records value will match the node's metadata key-value
pairs according to [RFC1464](https://www.ietf.org/rfc/rfc1464.txt).
Alternatively, the TXT record will only include the node's metadata value when the
node's metadata key starts with `rfc1035-`.
## Service Lookups
A service lookup is used to query for service providers. Service queries support
two lookup methods: standard and strict [RFC 2782](https://tools.ietf.org/html/rfc2782).
By default, SRV weights are all set at 1, but changing weights is supported using the
`Weights` attribute of the [service definition](/docs/agent/services).
Note that DNS is limited in size per request, even when performing DNS TCP
queries.
For services having many instances (more than 500), it might not be possible to
retrieve the complete list of instances for the service.
When DNS SRV response are sent, order is randomized, but weights are not
taken into account. In the case of truncation different clients using weighted SRV
responses will have partial and inconsistent views of instances weights so the
request distribution could be skewed from the intended weights. In that case,
it is recommended to use the HTTP API to retrieve the list of nodes.
### Standard Lookup
The format of a standard service lookup is:
```text
[tag.]<service>.service[.datacenter].<domain>
```
The `tag` is optional, and, as with node lookups, the `datacenter` is as
well. If no tag is provided, no filtering is done on tag. If no
datacenter is provided, the datacenter of this Consul agent is assumed.
If we want to find any redis service providers in our local datacenter,
we could query `redis.service.consul.` If we want to find the PostgreSQL
primary in a particular datacenter, we could query
`primary.postgresql.service.dc2.consul.`
The DNS query system makes use of health check information to prevent routing
to unhealthy nodes. When a service query is made, any services failing their health
check or failing a node system check will be omitted from the results. To allow
for simple load balancing, the set of nodes returned is also randomized each time.
These mechanisms make it easy to use DNS along with application-level retries
as the foundation for an auto-healing service oriented architecture.
For standard services queries, both A and SRV records are supported. SRV records
provide the port that a service is registered on, enabling clients to avoid relying
on well-known ports. SRV records are only served if the client specifically requests
them, like so:
```shell-session
$ dig @127.0.0.1 -p 8600 consul.service.consul SRV
; <<>> DiG 9.8.3-P1 <<>> @127.0.0.1 -p 8600 consul.service.consul ANY
; (1 server found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 50483
;; flags: qr aa rd; QUERY: 1, ANSWER: 3, AUTHORITY: 1, ADDITIONAL: 1
;; WARNING: recursion requested but not available
;; QUESTION SECTION:
;consul.service.consul. IN SRV
;; ANSWER SECTION:
consul.service.consul. 0 IN SRV 1 1 8300 foobar.node.dc1.consul.
;; ADDITIONAL SECTION:
foobar.node.dc1.consul. 0 IN A 10.1.10.12
```
### RFC 2782 Lookup
The format for RFC 2782 SRV lookups is:
_<service>._<protocol>[.service][.datacenter][.domain]
Per [RFC 2782](https://tools.ietf.org/html/rfc2782), SRV queries should use
underscores, `_`, as a prefix to the `service` and `protocol` values in a query to
prevent DNS collisions. The `protocol` value can be any of the tags for a
service. If the service has no tags, `tcp` should be used. If `tcp`
is specified as the protocol, the query will not perform any tag filtering.
Other than the query format and default `tcp` protocol/tag value, the behavior
of the RFC style lookup is the same as the standard style of lookup.
If you registered the service `rabbitmq` on port 5672 and tagged it with `amqp`,
you could make an RFC 2782 query for its SRV record as `_rabbitmq._amqp.service.consul`:
```shell-session
$ dig @127.0.0.1 -p 8600 _rabbitmq._amqp.service.consul SRV
; <<>> DiG 9.8.3-P1 <<>> @127.0.0.1 -p 8600 _rabbitmq._amqp.service.consul ANY
; (1 server found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 52838
;; flags: qr aa rd; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1
;; WARNING: recursion requested but not available
;; QUESTION SECTION:
;_rabbitmq._amqp.service.consul. IN SRV
;; ANSWER SECTION:
_rabbitmq._amqp.service.consul. 0 IN SRV 1 1 5672 rabbitmq.node1.dc1.consul.
;; ADDITIONAL SECTION:
rabbitmq.node1.dc1.consul. 0 IN A 10.1.11.20
```
Again, note that the SRV record returns the port of the service as well as its IP.
### Prepared Query Lookups
The format of a prepared query lookup is:
```text
<query or name>.query[.datacenter].<domain>
```
The `datacenter` is optional, and if not provided, the datacenter of this Consul
agent is assumed.
The `query or name` is the ID or given name of an existing
[Prepared Query](/api/query). These behave like standard service
queries but provide a much richer set of features, such as filtering by multiple
tags and automatically failing over to look for services in remote datacenters if
no healthy nodes are available in the local datacenter. Consul 0.6.4 and later also
added support for [prepared query templates](/api/query#prepared-query-templates)
which can match names using a prefix match, allowing one template to apply to
potentially many services.
To allow for simple load balancing, the set of nodes returned is randomized each time.
Both A and SRV records are supported. SRV records provide the port that a service is
registered on, enabling clients to avoid relying on well-known ports. SRV records are
only served if the client specifically requests them.
### Connect-Capable Service Lookups
To find Connect-capable services:
```text
<service>.connect.<domain>
```
This will find all [Connect-capable](/docs/connect)
endpoints for the given `service`. A Connect-capable endpoint may be
both a proxy for a service or a natively integrated Connect application.
The DNS interface does not differentiate the two.
Most services will use a [proxy](/docs/connect/proxies) that handles
service discovery automatically and therefore won't use this DNS format.
This DNS format is primarily useful for [Connect-native](/docs/connect/native)
applications.
This endpoint currently only finds services within the same datacenter
and doesn't support tags. This DNS interface will be expanded over time.
If you need more complex behavior, please use the
[catalog API](/api/catalog).
### Ingress Service Lookups
To find ingress-enabled services:
```text
<service>.ingress.<domain>
```
This will find all [ingress gateway](/docs/connect/ingress-gateway)
endpoints for the given `service`.
This endpoint currently only finds services within the same datacenter
and doesn't support tags. This DNS interface will be expanded over time.
If you need more complex behavior, please use the
[catalog API](/api/catalog).
### UDP Based DNS Queries
When the DNS query is performed using UDP, Consul will truncate the results
without setting the truncate bit. This is to prevent a redundant lookup over
TCP that generates additional load. If the lookup is done over TCP, the results
are not truncated.
## Caching
By default, all DNS results served by Consul set a 0 TTL value. This disables
caching of DNS results. However, there are many situations in which caching is
desirable for performance and scalability. This is discussed more in the tutorial
for [DNS caching](https://learn.hashicorp.com/tutorials/consul/dns-caching).
## WAN Address Translation
By default, Consul DNS queries will return a node's local address, even when
being queried from a remote datacenter. If you need to use a different address
to reach a node from outside its datacenter, you can configure this behavior
using the [`advertise-wan`](/docs/agent/options#_advertise-wan) and
[`translate_wan_addrs`](/docs/agent/options#translate_wan_addrs) configuration
options.
## Namespaced Services <EnterpriseAlert inline />
Consul Enterprise 1.7.0 added support for namespaces including resolving namespaced
services via DNS. To maintain backwards compatibility existing queries can be used
and these will resolve services within the `default` namespace. However, for resolving
services from other namespaces the following form can be used:
```text
[tag.]<service>.service.<namespace>.<datacenter>.<domain>
```
This is the canonical name of a Consul Enterprise service with all parts present. Like
Consul OSS some parts may be omitted but which parts depend on the value of the
[`prefer_namespace` configuration](/docs/agent/options#dns_prefer_namespace).
With `prefer_namespace` set to `true` the datacenter may be omitted and will be defaulted
to the local agents datacenter:
```text
[tag.]<service>.service.<namespace>.<domain>
```
With `prefer_namespace` set to `false` the namespace may be omitted and will be defaulted
to the `default` namespace:
```text
[tag.]<service>.service.<datacenter>
```
Finally, both the namespace and datacenter may be omitted and the service will be resolved
in the `default` namespace and in the datacenter of the local agent.
## DNS with ACLs
In order to use the DNS interface when
[Access Control Lists (ACLs)](/docs/security/acl/acl-system)
are enabled, you must first create ACL tokens with the necessary policies.
Consul agents resolve DNS requests using one of the preconfigured tokens below,
listed in order of precedence:
1. The agent's [`default` token](/docs/agent/options#acl_tokens_default).
2. The built-in [`anonymous` token](/docs/security/acl/acl-system#builtin-tokens).
Because the anonymous token is used when any request is made to Consul without
explicitly specifying a token, production deployments should not apply policies
needed for DNS to this token.
Consul will either accept or deny the request depending on whether the token
has the appropriate authorization. The following table describes the available
DNS lookups and required policies when ACLs are enabled:
| Lookup | Type | Description | ACLs Required |
| ---------------------------------------------------------- | -------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `*.node.consul` | [Node](#node-lookups) | Allow resolving DNS requests for the target node (i.e., `<target>.node.consul`) | [`node:read`](/docs/security/acl/acl-rules#node-rules) |
| `*.service.consul`, `*.connect.consul`, `*.ingress.consul` | [Service: standard](#service-lookups) | Allow resolving DNS requests for target service (e.g., `<target>.service.consul`) instances running on ACL-authorized nodes | [`service:read`](/docs/security/acl/acl-rules#service-rules), [`node:read`](/docs/security/acl/acl-rules#node-rules) |
| `*.query.consul` | [Service: prepared query](#prepared-query-lookups) | Allow resolving DNS requests for [service instances specified](/api/query#service-1) by the target prepared query (i.e., `<target>.query.consul`) running on ACL-authorized nodes | [`query:read`](/docs/security/acl/acl-rules#prepared-query-rules), [`service:read`](/docs/security/acl/acl-rules#service-rules), [`node:read`](/docs/security/acl/acl-rules#node-rules) |
For guidance on how to configure an appropriate token for DNS, refer to the
securing Consul with ACLs guides for:
- [Production Environments](https://learn.hashicorp.com/tutorials/consul/access-control-setup-production#token-for-dns)
- [Development Environments](https://learn.hashicorp.com/tutorials/consul/access-control-setup#additional-acl-configuration)