Documentation for version v1.10.1 is no longer actively maintained. The version you are currently viewing is a static snapshot. For up-to-date documentation, see the latest version.
Annotations are used in Ingress Controllers to configure features that are not covered by the Kubernetes Ingress API.
Some of the features that have been historically configured via annotations are supported as first-class features in Contour’s HTTPProxy API, which provides a more robust configuration interface over annotations.
However, Contour still supports a number of annotations on the Ingress resources.
The following Kubernetes annotations are supported on Ingress
objects:
The Ingress class annotation can be used to specify which Ingress controller should serve a particular Ingress object.
This annotation may be specified as the standard kubernetes.io/ingress.class
or a Contour-specific projectcontour.io/ingress.class
.
In both cases, they will behave as follows, by default:
kubernetes.io/ingress.class: contour
, then Contour serves the Ingress.You can override the default class contour
by providing the --ingress-class-name
flag to Contour.
This can be useful while you are migrating from another controller, or if you need multiple instances of Contour.
If you do this, the behavior is as follows:
--ingress-class-name
flag, Contour will ignore the Ingress.--ingress-class-name
flag, Contour will serve the Ingress.This same logic applies for these annotations on HTTPProxy objects.
ingress.kubernetes.io/force-ssl-redirect
: Requires TLS/SSL for the Ingress to Envoy by setting the Envoy virtual host option require_tls.kubernetes.io/ingress.allow-http
: Instructs Contour to not create an Envoy HTTP route for the virtual host. The Ingress exists only for HTTPS requests. Specify "false"
for Envoy to mark the endpoint as HTTPS only. All other values are ignored.The ingress.kubernetes.io/force-ssl-redirect
annotation takes precedence over kubernetes.io/ingress.allow-http
. If they are set to "true"
and "false"
respectively, Contour will create an Envoy HTTP route for the Virtual host, and set the require_tls
virtual host option.
projectcontour.io/ingress.class
: The Ingress class that should interpret and serve the Ingress. See the main Ingress class annotation section for more details.projectcontour.io/num-retries
: The maximum number of retries Envoy should make before abandoning and returning an error to the client. Applies only if projectcontour.io/retry-on
is specified.projectcontour.io/per-try-timeout
: The timeout per retry attempt, if there should be one. Applies only if projectcontour.io/retry-on
is specified.projectcontour.io/response-timeout
: The Envoy HTTP route timeout, specified as a golang duration. By default, Envoy has a 15 second timeout for a backend service to respond. Set this to infinity
to specify that Envoy should never timeout the connection to the backend. Note that the value 0s
/ zero has special semantics for Envoy.projectcontour.io/retry-on
: The conditions for Envoy to retry a request. See also possible values and their meanings for retry-on
.projectcontour.io/tls-minimum-protocol-version
: The minimum TLS protocol version the TLS listener should support. Valid options are 1.3
, 1.2
(default), 1.1
.projectcontour.io/websocket-routes
: The routes supporting websocket protocol, the annotation value contains a list of route paths separated by a comma that must match with the ones defined in the Ingress
definition. Defaults to Envoy’s default behavior which is use_websocket
to false
.A Kubernetes Service maps to an Envoy Cluster. Envoy clusters have many settings to control specific behaviors. These annotations allow access to some of those settings.
projectcontour.io/max-connections
: The maximum number of connections that a single Envoy instance allows to the Kubernetes Service; defaults to 1024.projectcontour.io/max-pending-requests
: The maximum number of pending requests that a single Envoy instance allows to the Kubernetes Service; defaults to 1024.projectcontour.io/max-requests
: The maximum parallel requests a single Envoy instance allows to the Kubernetes Service; defaults to 1024projectcontour.io/max-retries
: The maximum number of parallel retries a single Envoy instance allows to the Kubernetes Service; defaults to 1024. This is independent of the per-Kubernetes Ingress number of retries (projectcontour.io/num-retries
) and retry-on (projectcontour.io/retry-on
), which control whether retries are attempted and how many times a single request can retry.projectcontour.io/upstream-protocol.{protocol}
: The protocol used to proxy requests to the upstream service.
The annotation value contains a comma-separated list of port names and/or numbers that must match with the ones defined in the Service
definition.
This value can also be specified in the spec.routes.services[].protocol
field on the HTTPProxy object, where it takes precedence over the Service annotation.
Supported protocol names are: h2
, h2c
, and tls
:
tls
protocol allows for requests which terminate at Envoy to proxy via TLS to the upstream.
This protocol should be used for HTTP/1.1 services over TLS.
Note that validating the upstream TLS certificate requires additionally setting the validation field.h2
protocol proxies requests to the upstream using HTTP/2 over TLS.h2c
protocol proxies requests to the the upstream using cleartext HTTP/2.projectcontour.io/ingress.class
: The Ingress class that should interpret and serve the HTTPProxy. See the main Ingress class annotation section for more details.Read our getting started documentation.