NGINX Documentation

VirtualServer and VirtualServerRoute Resources

The VirtualServer and VirtualServerRoute resources are new load balancing configuration, introduced in release 1.5 as an alternative to the Ingress resource. The resources enable use cases not supported with the Ingress resource, such as traffic splitting and advanced content-based routing. The resources are implemented as Custom Resources.

This document is the reference documentation for the resources. To see additional examples of using the resources for specific use cases, go to the examples-of-custom-resources folder in our GitHub repo.

Contents

VirtualServer Specification

The VirtualServer resource defines load balancing configuration for a domain name, such as example.com. Below is an example of such configuration:

apiVersion: k8s.nginx.org/v1
kind: VirtualServer
metadata:
  name: cafe
spec:
  host: cafe.example.com
  tls:
    secret: cafe-secret
  upstreams:
  - name: tea
    service: tea-svc
    port: 80
  - name: coffee
    service: coffee-svc
    port: 80
  routes:
  - path: /tea
    action:
      pass: tea
  - path: /coffee
    action:
      pass: coffee
  - path: ~ ^/decaf/.*\\.jpg$
    action:
      pass: coffee
  - path: = /green/tea
    action:
      pass: tea
Field Description Type Required
host The host (domain name) of the server. Must be a valid subdomain as defined in RFC 1123, such as my-app or hello.example.com. Wildcard domains like *.example.com are not allowed. string Yes
tls The TLS termination configuration. tls No
upstreams A list of upstreams. []upstream No
routes A list of routes. []route No

VirtualServer.TLS

The tls field defines TLS configuration for a VirtualServer. For example:

secret: cafe-secret
redirect:
  enable: true
Field Description Type Required
secret The name of a secret with a TLS certificate and key. The secret must belong to the same namespace as the VirtualServer. The secret must contain keys named tls.crt and tls.key that contain the certificate and private key as described here. If the secret doesn’t exist, NGINX will break any attempt to establish a TLS connection to the host of the VirtualServer. string No
redirect The redirect configuration of the TLS for a VirtualServer. tls.redirect No

VirtualServer.TLS.Redirect

The redirect field configures a TLS redirect for a VirtualServer:

enable: true
code: 301
basedOn: scheme
Field Description Type Required
enable Enables a TLS redirect for a VirtualServer. The default is False. boolean No
code The status code of a redirect. The allowed values are: 301, 302, 307, 308. The default is 301. int No
basedOn The attribute of a request that NGINX will evaluate to send a redirect. The allowed values are scheme (the scheme of the request) or x-forwarded-proto (the X-Forwarded-Proto header of the request). The default is scheme. string No

VirtualServer.Route

The route defines rules for matching client requests to actions like passing a request to an upstream. For example:

  path: /tea
  action:
    pass: tea
Field Description Type Required
path The path of the route. NGINX will match it against the URI of a request. Possible values are: a prefix (/, /path), an exact match (=/exact/match), a case insensitive regular expression (~*^/Bar.*\\.jpg) or a case sensitive regular expression (~^/foo.*\\.jpg). In the case of a prefix (must start with /) or an exact match (must start with =), the path must not include any whitespace characters, {, } or ;. In the case of the regex matches, all double quotes " must be escaped and the match can’t end in an unescaped backslash \. The path must be unique among the paths of all routes of the VirtualServer. Check the location directive for more information. string Yes
action The default action to perform for a request. action No*
splits The default splits configuration for traffic splitting. Must include at least 2 splits. []split No*
matches The matching rules for advanced content-based routing. Requires the default action or splits. Unmatched requests will be handled by the default action or splits. matches No
route The name of a VirtualServerRoute resource that defines this route. If the VirtualServerRoute belongs to a different namespace than the VirtualServer, you need to include the namespace. For example, tea-namespace/tea. string No*

* – a route must include exactly one of the following: action, splits, or route.

VirtualServerRoute Specification

The VirtualServerRoute resource defines a route for a VirtualServer. It can consist of one or multiple subroutes. The VirtualServerRoute is an alternative to Mergeable Ingress types.

In the example below, the VirtualServer cafe from the namespace cafe-ns defines a route with the path /coffee, which is further defined in the VirtualServerRoute coffee from the namespace coffee-ns.

VirtualServer:

apiVersion: k8s.nginx.org/v1
kind: VirtualServer
metadata:
  name: cafe
  namespace: cafe-ns
spec:
  host: cafe.example.com
  upstreams:
  - name: tea
    service: tea-svc
    port: 80
  routes:
  - path: /tea
    action:
      pass: tea
  - path: /coffee
    route: coffee-ns/coffee

VirtualServerRoute:

apiVersion: k8s.nginx.org/v1
kind: VirtualServerRoute
metadata:
  name: coffee
  namespace: coffee-ns
spec:
  host: cafe.example.com
  upstreams:
  - name: latte
    service: latte-svc
    port: 80
  - name: espresso
    service: espresso-svc
    port: 80
  subroutes:
  - path: /coffee/latte
    action:
      pass: latte
  - path: /coffee/espresso
    action:
      pass: espresso

Note that each subroute must have a path that starts with the same prefix (here /coffee), which is defined in the route of the VirtualServer. Additionally, the host in the VirtualServerRoute must be the same as the host of the VirtualServer.

Field Description Type Required
host The host (domain name) of the server. Must be a valid subdomain as defined in RFC 1123, such as my-app or hello.example.com. Wildcard domains like *.example.com are not allowed. Must be the same as the host of the VirtualServer that references this resource. string Yes
upstreams A list of upstreams. []upstream No
subroutes A list of subroutes. []subroute No

VirtualServerRoute.Subroute

The subroute defines rules for matching client requests to actions like passing a request to an upstream. For example:

path: /coffee
action:
  pass: coffee
Field Description Type Required
path The path of the subroute. NGINX will match it against the URI of a request. Possible values are: a prefix (/, /path), an exact match (=/exact/match), a case insensitive regular expression (~*^/Bar.*\\.jpg) or a case sensitive regular expression (~^/foo.*\\.jpg). In the case of a prefix, the path must start with the same path as the path of the route of the VirtualServer that references this resource. In the case of an exact or regex match, the path must be the same as the path of the route of the VirtualServer that references this resource. In the case of a prefix or an exact match, the path must not include any whitespace characters, {, } or ;. In the case of the regex matches, all double quotes " must be escaped and the match can’t end in an unescaped backslash \. The path must be unique among the paths of all subroutes of the VirtualServerRoute. string Yes
action The default action to perform for a request. action No*
splits The default splits configuration for traffic splitting. Must include at least 2 splits. []split No*
matches The matching rules for advanced content-based routing. Requires the default action or splits. Unmatched requests will be handled by the default action or splits. matches No

* – a subroute must include exactly one of the following: action or splits.

Common Parts of the VirtualServer and VirtualServerRoute

Upstream

The upstream defines a destination for the routing configuration. For example:

name: tea
service: tea-svc
subselector:
  version: canary
port: 80
lb-method: round_robin
fail-timeout: 10s
max-fails: 1
max-conns: 32
keepalive: 32
connect-timeout: 30s
read-timeout: 30s
send-timeout: 30s
next-upstream: "error timeout non_idempotent"
next-upstream-timeout: 5s
next-upstream-tries: 10
client-max-body-size: 2m
tls:
  enable: true

Note: The WebSocket protocol is supported without any additional configuration.

Field Description Type Required
name The name of the upstream. Must be a valid DNS label as defined in RFC 1035. For example, hello and upstream-123 are valid. The name must be unique among all upstreams of the resource. string Yes
service The name of a service. The service must belong to the same namespace as the resource. If the service doesn’t exist, NGINX will assume the service has zero endpoints and return a 502 response for requests for this upstream. For NGINX Plus only, services of type ExternalName are also supported (check the prerequisites). string Yes
subselector Selects the pods within the service using label keys and values. By default, all pods of the service are selected. Note: the specified labels are expected to be present in the pods when they are created. If the pod labels are updated, the Ingress Controller will not see that change until the number of the pods is changed. map[string]string No
port The port of the service. If the service doesn’t define that port, NGINX will assume the service has zero endpoints and return a 502 response for requests for this upstream. The port must fall into the range 1..65553. uint16 Yes
lb-method The load balancing method. To use the round-robin method, specify round_robin. The default is specified in the lb-method ConfigMap key. string No
fail-timeout The time during which the specified number of unsuccessful attempts to communicate with an upstream server should happen to consider the server unavailable. See the fail_timeout parameter of the server directive. The default is set in the fail-timeout ConfigMap key. string No
max-fails The number of unsuccessful attempts to communicate with an upstream server that should happen in the duration set by the fail-timeout to consider the server unavailable. See the max_fails parameter of the server directive. The default is set in the max-fails ConfigMap key. int No
max-conns The maximum number of simultaneous active connections to an upstream server. See the max_conns parameter of the server directive. By default there is no limit. Note: if keepalive connections are enabled, the total number of active and idle keepalive connections to an upstream server may exceed the max_conns value. int No
keepalive Configures the cache for connections to upstream servers. The value 0 disables the cache. See the keepalive directive. The default is set in the keepalive ConfigMap key. int No
connect-timeout The timeout for establishing a connection with an upstream server. See the proxy_connect_timeout directive. The default is specified in the proxy-connect-timeout ConfigMap key. string No
read-timeout The timeout for reading a response from an upstream server. See the proxy_read_timeout directive. The default is specified in the proxy-read-timeout ConfigMap key. string No
send-timeout The timeout for transmitting a request to an upstream server. See the proxy_send_timeout directive. The default is specified in the proxy-send-timeout ConfigMap key. string No
next-upstream Specifies in which cases a request should be passed to the next upstream server. See the proxy_next_upstream directive. The default is error timeout. string No
next-upstream-timeout The time during which a request can be passed to the next upstream server. See the proxy_next_upstream_timeout directive. The 0 value turns off the time limit. The default is 0. string No
next-upstream-tries The number of possible tries for passing a request to the next upstream server. See the proxy_next_upstream_tries directive. The 0 value turns off this limit. The default is 0. int No
client-max-body-size Sets the maximum allowed size of the client request body. See the client_max_body_size directive. The default is set in the client-max-body-size ConfigMap key. string No
tls The TLS configuration for the Upstream. tls No
healthCheck The health check configuration for the Upstream. See the health_check directive. Note: this feature is supported only in NGINX Plus. healthcheck No
slow-start The slow start allows an upstream server to gradually recover its weight from 0 to its nominal value after it has been recovered or became available or when the server becomes available after a period of time it was considered unavailable. By default, the slow start is disabled. See the slow_start parameter of the server directive. Note: The parameter cannot be used along with the random, hash or ip_hash load balancing methods and will be ignored. string No
queue Configures a queue for an upstream. A client request will be placed into the queue if an upstream server cannot be selected immediately while processing the request. By default, no queue is configured. Note: this feature is supported only in NGINX Plus. queue No
buffering Enables buffering of responses from the upstream server. See the proxy_buffering directive. The default is set in the proxy-buffering ConfigMap key. boolean No
buffers Configures the buffers used for reading a response from the upstream server for a single connection. buffers No
buffer-size Sets the size of the buffer used for reading the first part of a response received from the upstream server. See the proxy_buffer_size directive. The default is set in the proxy-buffer-size ConfigMap key. string No

Upstream.Buffers

The buffers field configures the buffers used for reading a response from the upstream server for a single connection:

number: 4
size: 8K

See the proxy_buffers directive for additional information.

Field Description Type Required
number Configures the number of buffers. The default is set in the proxy-buffers ConfigMap key. int Yes
size Configures the size of a buffer. The default is set in the proxy-buffers ConfigMap key. string Yes

Upstream.TLS

Field Description Type Required
enable Enables HTTPS for requests to upstream servers. The default is False, meaning that HTTP will be used. boolean No

Upstream.Queue

The queue field configures a queue. A client request will be placed into the queue if an upstream server cannot be selected immediately while processing the request:

size: 10
timeout: 60s

See queue directive for additional information.

Note: This feature is supported only in NGINX Plus.

Field Description Type Required
size The size of the queue. int Yes
timeout The timeout of the queue. A request cannot be queued for a period longer than the timeout. The default is 60s. string No

Upstream.Healthcheck

The Healthcheck defines an active health check. In the example below we enable a health check for an upstream and configure all the available parameters:

name: tea
service: tea-svc
port: 80
healthCheck:
  enable: true
  path: /healthz
  interval: 20s
  jitter: 3s
  fails: 5
  passes: 5
  port: 8080
  tls:
    enable: true
  connect-timeout: 10s
  read-timeout: 10s
  send-timeout: 10s
  headers:
  - name: Host
    value: my.service
  statusMatch: "! 500"
Field Description Type Required
enable Enables a health check for an upstream server. The default is false. boolean No
path The path used for health check requests. The default is /. string No
interval The interval between two consecutive health checks. The default is 5s. string No
jitter The time within which each health check will be randomly delayed. By default, there is no delay. string No
fails The number of consecutive failed health checks of a particular upstream server after which this server will be considered unhealthy. The default is 1. integer No
passes The number of consecutive passed health checks of a particular upstream server after which the server will be considered healthy. The default is 1. integer No
port The port used for health check requests. By default, the port of the upstream is used. Note: in contrast with the port of the upstream, this port is not a service port, but a port of a pod. integer No
tls The TLS configuration used for health check requests. By default, the tls field of the upstream is used. upstream.tls No
connect-timeout The timeout for establishing a connection with an upstream server. By default, the connect-timeout of the upstream is used. string No
read-timeout The timeout for reading a response from an upstream server. By default, the read-timeout of the upstream is used. string No
send-timeout The timeout for transmitting a request to an upstream server. By default, the send-timeout of the upstream is used. string No
headers The request headers used for health check requests. NGINX Plus always sets the Host, User-Agent and Connection headers for health check requests. []header No
statusMatch The expected response status codes of a health check. By default, the response should have status code 2xx or 3xx. Examples: "200", "! 500", "301-303 307". See the documentation of the match directive. string No

Upstream.SessionCookie

The SessionCookie field configures session persistence which allows requests from the same client to be passed to the same upstream server. The information about the designated upstream server is passed in a session cookie generated by NGINX Plus.

In the example below, we configure session persistence with a session cookie for an upstream and configure all the available parameters:

name: tea
service: tea-svc
port: 80
sessionCookie:
  enable: true
  name: srv_id
  path: /
  expires: 1h
  domain: .example.com
  httpOnly: false
  secure: true

See the sticky directive for additional information. The session cookie corresponds to the sticky cookie method.

Field Description Type Required
enable Enables session persistence with a session cookie for an upstream server. The default is false. boolean No
name The name of the cookie. string Yes
path The path for which the cookie is set. string No
expires The time for which a browser should keep the cookie. Can be set to the special value max, which will cause the cookie to expire on 31 Dec 2037 23:55:55 GMT. string No
domain The domain for which the cookie is set. string No
httpOnly Adds the HttpOnly attribute to the cookie. boolean No
secure Adds the Secure attribute to the cookie. boolean No

Action

The action defines an action to perform for a request.

In the example below, client requests are passed to an upstream coffee:

 path: /coffee
 action:
  pass: coffee
Field Description Type Required
pass Passes requests to an upstream. The upstream with that name must be defined in the resource. string No*
redirect Redirects requests to a provided URL. action.redirect No*
return Returns a preconfigured response. action.return No*

* – an action must include exactly one of the following: pass, redirect or return.

Action.Redirect

The redirect action defines a redirect to return for a request.

In the example below, client requests are passed to a url http://www.nginx.com:

redirect:
  url: http://www.nginx.com
  code: 301
Field Description Type Required
url The URL to redirect the request to. Supported NGINX variables: $scheme, $http_x_forwarded_proto, $request_uri, $host. Variables must be inclosed in curly braces. For example: ${host}${request_uri}. string Yes
code The status code of a redirect. The allowed values are: 301, 302, 307, 308. The default is 301. int No

Action.Return

The return action defines a preconfigured response for a request.

In the example below, NGINX will respond with the preconfigured response for every request:

return:
  code: 200
  type: text/plain
  body: "Hello World\n"
Field Description Type Required
code The status code of the response. The allowed values are: 2XX, 4XX or 5XX. The default is 200. int No
type The MIME type of the response. The default is text/plain. string No
body The body of the response. Supports NGINX variables*. Variables must be inclosed in curly brackets. For example: Request is ${request_uri}\n. string Yes

* – Supported NGINX variables: $request_uri, $request_method, $request_body, $scheme, $http_, $args, $arg_, $cookie_, $host, $request_time, $request_length, $nginx_version, $pid, $connection, $remote_addr, $remote_port, $time_iso8601, $time_local, $server_addr, $server_port, $server_name, $server_protocol, $connections_active, $connections_reading, $connections_writing and $connections_waiting.

Split

The split defines a weight for an action as part of the splits configuration.

In the example below NGINX passes 80% of requests to the upstream coffee-v1 and the remaining 20% to coffee-v2:

splits:
- weight: 80
  action:
    pass: coffee-v1
- weight: 20
  action:
    pass: coffee-v2
Field Description Type Required
weight The weight of an action. Must fall into the range 1..99. The sum of the weights of all splits must be equal to 100. int Yes
action The action to perform for a request. action Yes

Match

The match defines a match between conditions and an action or splits.

In the example below, NGINX routes requests with the path /coffee to different upstreams based on the value of the cookie user:

  • user=john -> coffee-future
  • user=bob -> coffee-deprecated
  • If the cookie is not set or not equal to either john or bob, NGINX routes to coffee-stable
path: /coffee
matches:
- conditions:
  - cookie: user
    value: john
  action:
    pass: coffee-future
- conditions:
  - cookie: user
    value: bob
  action:
    pass: coffee-deprecated
action:
  pass: coffee-stable

In the next example, NGINX routes requests based on the value of the built-in $request_method variable, which represents the HTTP method of a request:

  • all POST requests -> coffee-post
  • all non-POST requests -> coffee
path: /coffee
matches:
- conditions:
  - variable: $request_method
    value: POST
  action:
    pass: coffee-post
action:
  pass: coffee
Field Description Type Required
conditions A list of conditions. Must include at least 1 condition. []condition Yes
action The action to perform for a request. action No*
splits The splits configuration for traffic splitting. Must include at least 2 splits. []split No*

* – a match must include exactly one of the following: action or splits.

Condition

The condition defines a condition in a match.

Field Description Type Required
header The name of a header. Must consist of alphanumeric characters or -. string No*
cookie The name of a cookie. Must consist of alphanumeric characters or _. string No*
argument The name of an argument. Must consist of alphanumeric characters or _. string No*
variable The name of an NGINX variable. Must start with $. See the list of the supported variables below the table. string No*
value The value to match the condition against. How to define a value is shown below the table. string Yes

* – a condition must include exactly one of the following: header, cookie, argument or variable.

Supported NGINX variables: $args, $http2, $https, $remote_addr, $remote_port, $query_string, $request, $request_body, $request_uri, $request_method, $scheme. Find the documentation for each variable here.

The value supports two kinds of matching:

  • Case-insensitive string comparison. For example:
    • john – case-insensitive matching that succeeds for strings, such as john, John, JOHN.
    • !john – negation of the case-incentive matching for john that succeeds for strings, such as bob, anything, '' (empty string).
  • Matching with a regular expression. Note that NGINX supports regular expressions compatible with those used by the Perl programming language (PCRE). For example:
    • ~^yes – a case-sensitive regular expression that matches any string that starts with yes. For example: yes, yes123.
    • !~^yes – negation of the previous regular expression that succeeds for strings like YES, Yes123, noyes. (The negation mechanism is not part of the PCRE syntax).
    • ~*no$ – a case-insensitive regular expression that matches any string that ends with no. For example: no, 123no, 123NO.

Note: a value must not include any unescaped double quotes (") and must not end with an unescaped backslash (\). For example, the following are invalid values: some"value, somevalue\.

Using VirtualServer and VirtualServerRoute

You can use the usual kubectl commands to work with VirtualServer and VirtualServerRoute resources, similar to Ingress resources.

For example, the following command creates a VirtualServer resource defined in cafe-virtual-server.yaml with the name cafe:

$ kubectl apply -f cafe-virtual-server.yaml
virtualserver.k8s.nginx.org "cafe" created

You can get the resource by running:

$ kubectl get virtualserver cafe
NAME      AGE
cafe      3m

In the kubectl get and similar commands, you can also use the short name vs instead of virtualserver.

Working with VirtualServerRoute resources is analogous. In the kubectl commands, use virtualserverroute or the short name vsr.

Validation

The Ingress Controller validates VirtualServer and VirtualServerRoute resources. If a resource is invalid, the Ingress Controller will reject it.

You can check if the Ingress Controller successfully applied the configuration for a VirtualServer. For our example cafe VirtualServer, we can run:

$ kubectl describe vs cafe
. . .
Events:
  Type    Reason          Age   From                      Message
  ----    ------          ----  ----                      -------
  Normal  AddedOrUpdated  16s   nginx-ingress-controller  Configuration for default/cafe was added or updated

Note how the events section includes a Normal event with the AddedOrUpdated reason that informs us that the configuration was successfully applied.

If you create an invalid resource, the Ingress Controller will reject it and emit a Rejected event. For example, if you create a VirtualServer cafe with an empty host field, you will get:

$ kubectl describe vs cafe
. . .
Events:
  Type     Reason    Age   From                      Message
  ----     ------    ----  ----                      -------
  Warning  Rejected  2s    nginx-ingress-controller  VirtualServer default/cafe is invalid and was rejected: spec.host: Required value

Note how the events section includes a Warning event with the Rejected reason.

The Ingress Controller validates VirtualServerRoute resources in a similar way.

Note: If you make an existing resource invalid, the Ingress Controller will reject it and remove the corresponding configuration from NGINX.

Customization via ConfigMap

You can customize the NGINX configuration for VirtualServer and VirtualServerRoutes resources using the ConfigMap. Most of the ConfigMap keys are supported, with the following exceptions:

  • proxy-hide-headers
  • proxy-pass-headers
  • hsts
  • hsts-max-age
  • hsts-include-subdomains
  • hsts-behind-proxy
  • redirect-to-https
  • ssl-redirect