Skip to content

OSDOCS-13509: add assembly, topic map, and concept #92976

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

Merged
merged 1 commit into from
May 27, 2025
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
3 changes: 3 additions & 0 deletions _topic_maps/_topic_map.yml
Original file line number Diff line number Diff line change
Expand Up @@ -1705,6 +1705,9 @@ Topics:
- Name: Allocating load balancers to specific subnets
File: allocating-load-balancers
Distros: openshift-enterprise,openshift-origin
- Name: Gateway API with OpenShift Container Platform networking
File: ingress-gateway-api
Distros: openshift-enterprise
# Kubernetes NMState (TECHNOLOGY PREVIEW)
- Name: Kubernetes NMState
Dir: k8s_nmstate
Expand Down
138 changes: 138 additions & 0 deletions modules/nw-ingress-gateway-api-deployment-topologies.adoc
Original file line number Diff line number Diff line change
@@ -0,0 +1,138 @@
// Modules included in the following assemblies:
//
// * networking/gateway-api.adoc

:_mod-docs-content-type: CONCEPT
[id="nw-ingress-gateway-api-deployment_{context}"]
= Gateway API deployment topologies

Gateway API is designed to accomodate two topologies: shared gateways or dedicated gateways. Each topology has its own advantages and different security implications.

Dedicated gateway:: Routes and any load balancers or proxies are served from the same namespace. The `Gateway`
object restricts routes to a particular application namespace. This is the default topology when deploying a Gateway API resource in {product-title}.

Shared gateway:: Routes are served from multiple namespaces or with multiple hostnames. The `Gateway` object filters allow routes from application namespaces by using the `spec.listeners.allowedRoutes.namespaces` field.

[id="dedicated-gateway-example_{context}"]
== Dedicated gateway example
The following example shows a dedicated `Gateway` resource, `fin-gateway`:

.Example dedicated `Gateway` resource
[source,yaml]
----
apiVersion: gateway.networking.k8s.io/v1
kind: Gateway
metadata:
name: fin-gateway
namespace: openshift-ingress
spec:
listeners: <1>
- name: http
protocol: HTTP
port: 8080
hostname: "example.com"

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think the dedicated gateway can also serve multiple httproutes in Same namespace, so could use wildcard hostname like hostname: "*.example.com"
@alebedev87 Do you think we should consider the DNSRecords creation in the example, if yes then need to set hostname like *.dedicated.<baseDomain>, the baseDomin could be get from oc get dns.config cluster -ojsonpath="{.spec.baseDomain}"

Copy link
Contributor Author

@jldohmann jldohmann May 27, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

this sounds like it can be a continuous improvement task for a new PR. the current one is pretty large and has had enough scope creep already

----
<1> Creating a `Gateway` resource without setting `spec.listeners[].allowedRoutes` results in implicitly setting the `namespaces.from` field to have the value `Same`.

The following example shows the associated `HTTPRoute` resource, `sales-db`, which attaches to the dedicated `Gateway` object:

.Example `HTTPRoute` resource
[source,yaml]
----
apiVersion: gateway.networking.k8s.io/v1
kind: HTTPRoute
metadata:
name: sales-db
namespace: openshift-ingress

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

In this example the HTTPRoute sales-db is in openshift-ingress namespace, but it might not be a case in practice. We mentioned Dedicated gateway previously that

The `Gateway` object restricts routes to a particular application namespace.

Maybe we could add namespaces selector to gateway and label to namespace as well (similar to shared gateway but just apply to one dedicated namespace) ?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

i will make note of this for an updated PR with some continuous improvements

spec:
parentRefs:
- name: fin-gateway
hostnames:
- sales-db.example.com
rules:
- backendRefs:
- name: sales-db
¦ port: 8080
----

The `HTTPRoute` resource must have the name of the `Gateway` object as the value for its `parentRefs` field in order to attach to the gateway. Implicitly, the route is assumed to be in the same namespace as the `Gateway` object.

[id="shared-gateway-example_{context}"]
== Shared gateway example
The following example shows a `Gateway` resource, `devops-gateway`, that has a `spec.listeners.allowedRoutes.namespaces` label selector set to match any namespaces containing `shared-gateway-access: "true"`:

.Example shared `Gateway` resource
[source,yaml]
----
apiVersion: gateway.networking.k8s.io/v1
kind: Gateway
metadata:
name: devops-gateway
namespace: openshift-ingress
listeners:
- name: https
protocol: HTTPS
hostname: "example.com"
allowedRoutes:
namespaces:
from: Selector
selector:
¦ matchLabels:
¦ shared-gateway-access: "true"
----

The following examples show the allowed namespaces for the `devops-gateway` resource:

.Example `Namespace` resources
[source,yaml]
----
apiVersion: v1
kind: Namespace
metadata:
name: dev
labels:
shared-gateway-access: "true"
---
apiVersion: v1
kind: Namespace
metadata:
name: ops
labels:
shared-gateway-access: "true"
----

In this example, two `HTTPRoute` resources, `dev-portal` and `ops-home`, are in different namespaces but are attached to the shared gateway:

[source,yaml]
----
apiVersion: v1
kind: HTTPRoute
metadata:
name: dev-portal
namespace: dev
spec:
parentRefs:
- name: devops-gateway
namespace: openshift-ingress
rules:
- backendRefs:
- name: dev-portal
port: 8080
---
apiVersion: v1
kind: HTTPRoute
metadata:
name: ops-home
namespace: ops
spec:
parentRefs:
- name: devops-gateway
namespace: openshift-ingress
rules:
- backendRefs:
- name: ops-home
port: 8080
----

With a shared gateway topology, the routes must specify the namespace of the `Gateway` object it wants to attach to. Multiple `Gateway` objects can be deployed and shared across namespaces. When there are multiple shared gateways, this topology becomes conceptually similar to Ingress Controller sharding.

19 changes: 19 additions & 0 deletions modules/nw-ingress-gateway-api-implementation.adoc
Original file line number Diff line number Diff line change
@@ -0,0 +1,19 @@
// Modules included in the following assemblies:
//
// * networking/gateway-api.adoc

:_mod-docs-content-type: CONCEPT
[id="nw-ingress-gateway-api-implementation_{context}"]
= Gateway API implementation for {product-title}

The Ingress Operator manages the lifecycle of Gateway API CRDs in a way that enables other vendor implementations to make use of CRDs defined in an {product-title} cluster.

In some situations, the Gateway API provides one or more fields that a vendor implementation does not support, but that implementation is otherwise compatible in schema with the rest of the fields. These "dead fields" can result in disrupted Ingress workloads, improperly provisioned applications and services, and security related issues. Because {product-title} uses a specific version of Gateway API CRDs, any use of third-party implementations of Gateway API must conform to the {product-title} implementation to ensure that all fields work as expected.

Any CRDs created within an {product-title} {product-version} cluster are compatibly versioned and maintained by the Ingress Operator. If CRDs are already present but were not previously managed by the Ingress Operator, the Ingress Operator checks whether these configurations are compatible with the Gateway API version supported by {product-title}, and creates an admin-gate that requires your acknowledgment of CRD succession.

[IMPORTANT]
====
If you are updating your cluster from a previous {product-title} version that contains Gateway API CRDs change those resources so that they exactly match the version supported by {product-title}. Otherwise, you cannot update your cluster because those CRDs were not managed by {product-title}, and could contain functionality that is unsupported by Red{nbsp}Hat.
====

36 changes: 36 additions & 0 deletions modules/nw-ingress-gateway-api-overview.adoc
Original file line number Diff line number Diff line change
@@ -0,0 +1,36 @@
// Modules included in the following assemblies:
//
// * networking/gateway-api.adoc

:_mod-docs-content-type: CONCEPT
[id="nw-ingress-gateway-api-overview_{context}"]
= Overview of the Gateway API

The Gateway API is an open source, community-managed, Kubernetes networking mechanism. It focuses on routing within the transport layer, L4, and the application layer, L7, for clusters. A variety of vendors offer many link:https://gateway-api.sigs.k8s.io/implementations/[implementations of Gateway API].

The project is an effort to provide a standardized ecosystem by using a portable API with broad community support. By integrating Gateway API functionality into the Ingress Operator, it enables a networking solution that aligns with existing community and upstream development efforts.

The Gateway API extends the functionality of the Ingress Operator to handle more granular cluster traffic and routing configurations. With these capabilities, you can create instances of Gateway APIs custom resource definitions (CRDs). For {product-title} clusters, the Ingress Operator creates the following resources:

Gateway:: This resource describes how traffic can be translated to services within the cluster. For example, a specific load balancer configuration.
GatewayClass:: This resource defines a set of `Gateway` objects that share a common configuration and behavior. For example, two separate `GatewayClass` objects might be created to distinguish a set of `Gateway` resources used for public or private applications.
HTTPRoute:: This resource specifies the routing behavior of HTTP requests from a Gateway to a service, and is especially useful for multiplexing HTTP or terminated HTTPS connections.
GRPCRoute:: This resource specifies the routing behavior of gRPC requests.
ReferenceGrant:: This resource enables cross-namespace references. For example, it enables routes to forward traffic to backends that are in a different namespace.

In {product-title}, the implementation of the Gateway API is based on `gateway.networking.k8s.io/v1`, and all fields in this version are supported.

[id="gateway-api-benefits_{context}"]
== Benefits of the Gateway API
The Gateway API provides the following benefits:

* Portability: While {product-title} uses HAProxy to improve Ingress performance, Gateway API does not rely on vendor-specific annotations to provide certain behavior. To get comparable performance as HAProxy, the `Gateway` objects need to be horizontally scaled or their associated nodes need to be vertically scaled.
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

last sentence in this part ack'd by Shane in slack

* Separation of concerns: Gateway API uses a role-based approach to its resources, and more neatly fits into how a large organization structures its responsibilities and teams. Platform engineers might focus on `GatewayClass` resources, cluster admins might focus on configuring `Gateway` resources, and application developers might focus on routing their services with `HTTPRoute` resources.
* Extensibility: Additional functionality is developed as a standardized CRD.

[id="gateway-api-limitations_{context}"]
== Limitations of the Gateway API
The Gateway API has the following limitations:

* Version incompatibilites: The Gateway API ecosystem changes rapidly, and some implementations do not work with others because their featureset is based on differing versions of Gateway API.
* Resource overhead: While more flexible, Gateway API uses multiple resource types to achieve an outcome. For smaller applications, the simplicity of traditional Ingress might be a better fit.
Original file line number Diff line number Diff line change
@@ -0,0 +1,23 @@
:_mod-docs-content-type: ASSEMBLY
[id="ingress-gateway-api"]
= Gateway API with {product-title} Networking
include::_attributes/common-attributes.adoc[]
:context: ingress-gateway-api

toc::[]

{product-title} provides additional ways of configuring network traffic by using Gateway API with the Ingress Operator.

[IMPORTANT]
====
Gateway API does not support user-defined networks (UDN).
====

include::modules/nw-ingress-gateway-api-overview.adoc[leveloffset=+1]

include::modules/nw-ingress-gateway-api-implementation.adoc[leveloffset=+1]

include::modules/nw-ingress-gateway-api-deployment-topologies.adoc[leveloffset=+1]

.Additional resources
* xref:configuring-ingress-cluster-traffic-ingress-controller.adoc#nw-ingress-sharding-concept_configuring-ingress-cluster-traffic-ingress-controller[Ingress Controller sharding].