external-dns/source/ingress.go
Aleksei Sviridkin 5a55b09f48
feat(annotations): add custom annotation prefix support for split horizon DNS (#5889)
* feat(annotations): add custom annotation prefix support for split horizon DNS

Add --annotation-prefix flag to allow customizing the annotation prefix
used by external-dns. This enables split horizon DNS scenarios where
multiple instances process different sets of annotations from the same
Kubernetes resources.

Changes:
- Add AnnotationPrefix field to Config with validation
- Convert annotation constants to variables that can be reconfigured
- Add SetAnnotationPrefix() function to rebuild annotation keys
- Integrate annotation prefix setting in controller startup
- Update Helm chart with annotationPrefix value
- Add comprehensive split horizon DNS documentation
- Update FAQ with annotation prefix examples

This maintains full backward compatibility - the default prefix remains
"external-dns.alpha.kubernetes.io/".

Co-Authored-By: Claude <noreply@anthropic.com>

* docs(advanced): fix markdown formatting in split-horizon guide

Add blank lines before code blocks to improve markdown rendering
and comply with markdownlint rules.

Co-Authored-By: Claude <noreply@anthropic.com>

* docs(advanced): fix markdown formatting in split-horizon guide

Co-Authored-By: Claude <noreply@anthropic.com>

* docs(charts): regenerate Helm chart documentation

Co-Authored-By: Claude <noreply@anthropic.com>

* test: add AnnotationPrefix field to test configs

Add missing AnnotationPrefix field to minimalConfig and overriddenConfig
test configurations to match the new default value set in NewConfig().

Co-Authored-By: Claude <noreply@anthropic.com>

* test(charts): update error pattern in json-schema test

Update expected error message pattern to match current Helm validation
output format.

Co-Authored-By: Claude <noreply@anthropic.com>

* refactor(annotations): remove init() for explicit initialization

- Remove init() function from annotations package
- Add explicit SetAnnotationPrefix() call in controller/execute.go
- Remove annotation key aliases from source/source.go
- Replace all alias usages with annotations.* references (348 changes in 28 files)
- Add TestMain to existing test files (service_test.go, cloudflare_test.go)

This change makes annotation initialization explicit and predictable,
avoiding hidden global state initialization at import time.

Co-Authored-By: Claude <noreply@anthropic.com>

* docs: update changelog and mkdocs to include annotationPrefix and split horizon DNS

Signed-off-by: Aleksei Sviridkin <f@lex.la>

* docs(split-horizon): fix linting

Signed-off-by: Aleksei Sviridkin <f@lex.la>

* refactor(annotations): replace hardcoded annotation prefix with constant

Replace all hardcoded "external-dns.alpha.kubernetes.io/" strings
with annotations.DefaultAnnotationPrefix constant to establish
a single source of truth.

Changes:
- Add DefaultAnnotationPrefix constant in source/annotations/annotations.go
- Replace hardcoded string in controller/execute.go with constant reference
- Replace hardcoded strings in pkg/apis/externaldns/types.go (2 occurrences)
- Add helm unit tests for annotationPrefix value

This eliminates string duplication and makes future changes easier.

Co-Authored-By: Claude <noreply@anthropic.com>

---------

Signed-off-by: Aleksei Sviridkin <f@lex.la>
Co-authored-by: Claude <noreply@anthropic.com>
2025-11-08 03:56:52 -08:00

359 lines
12 KiB
Go

/*
Copyright 2017 The Kubernetes Authors.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
*/
package source
import (
"context"
"errors"
"fmt"
"sort"
"strings"
"text/template"
log "github.com/sirupsen/logrus"
networkv1 "k8s.io/api/networking/v1"
"k8s.io/apimachinery/pkg/labels"
"k8s.io/apimachinery/pkg/selection"
kubeinformers "k8s.io/client-go/informers"
netinformers "k8s.io/client-go/informers/networking/v1"
"k8s.io/client-go/kubernetes"
"sigs.k8s.io/external-dns/source/informers"
"sigs.k8s.io/external-dns/endpoint"
"sigs.k8s.io/external-dns/source/annotations"
"sigs.k8s.io/external-dns/source/fqdn"
)
const (
// Possible values for the ingress-hostname-source annotation
IngressHostnameSourceAnnotationOnlyValue = "annotation-only"
IngressHostnameSourceDefinedHostsOnlyValue = "defined-hosts-only"
IngressClassAnnotationKey = "kubernetes.io/ingress.class"
)
// ingressSource is an implementation of Source for Kubernetes ingress objects.
// Ingress implementation will use the spec.rules.host value for the hostname
// Use annotations.TargetKey to explicitly set Endpoint. (useful if the ingress
// controller does not update, or to override with alternative endpoint)
type ingressSource struct {
client kubernetes.Interface
namespace string
annotationFilter string
ingressClassNames []string
fqdnTemplate *template.Template
combineFQDNAnnotation bool
ignoreHostnameAnnotation bool
ingressInformer netinformers.IngressInformer
ignoreIngressTLSSpec bool
ignoreIngressRulesSpec bool
labelSelector labels.Selector
}
// NewIngressSource creates a new ingressSource with the given config.
func NewIngressSource(
ctx context.Context,
kubeClient kubernetes.Interface,
namespace, annotationFilter, fqdnTemplate string,
combineFqdnAnnotation, ignoreHostnameAnnotation, ignoreIngressTLSSpec, ignoreIngressRulesSpec bool,
labelSelector labels.Selector,
ingressClassNames []string) (Source, error) {
tmpl, err := fqdn.ParseTemplate(fqdnTemplate)
if err != nil {
return nil, err
}
// ensure that ingress class is only set in either the ingressClassNames or
// annotationFilter but not both
if ingressClassNames != nil && annotationFilter != "" {
selector, err := getLabelSelector(annotationFilter)
if err != nil {
return nil, err
}
requirements, _ := selector.Requirements()
for _, requirement := range requirements {
if requirement.Key() == "kubernetes.io/ingress.class" {
return nil, errors.New("--ingress-class is mutually exclusive with the kubernetes.io/ingress.class annotation filter")
}
}
}
// Use shared informer to listen for add/update/delete of ingresses in the specified namespace.
// Set resync period to 0, to prevent processing when nothing has changed.
informerFactory := kubeinformers.NewSharedInformerFactoryWithOptions(kubeClient, 0, kubeinformers.WithNamespace(namespace))
ingressInformer := informerFactory.Networking().V1().Ingresses()
// Add default resource event handlers to properly initialize informer.
_, _ = ingressInformer.Informer().AddEventHandler(informers.DefaultEventHandler())
informerFactory.Start(ctx.Done())
// wait for the local cache to be populated.
if err := informers.WaitForCacheSync(context.Background(), informerFactory); err != nil {
return nil, err
}
sc := &ingressSource{
client: kubeClient,
namespace: namespace,
annotationFilter: annotationFilter,
ingressClassNames: ingressClassNames,
fqdnTemplate: tmpl,
combineFQDNAnnotation: combineFqdnAnnotation,
ignoreHostnameAnnotation: ignoreHostnameAnnotation,
ingressInformer: ingressInformer,
ignoreIngressTLSSpec: ignoreIngressTLSSpec,
ignoreIngressRulesSpec: ignoreIngressRulesSpec,
labelSelector: labelSelector,
}
return sc, nil
}
// Endpoints returns endpoint objects for each host-target combination that should be processed.
// Retrieves all ingress resources on all namespaces
func (sc *ingressSource) Endpoints(_ context.Context) ([]*endpoint.Endpoint, error) {
ingresses, err := sc.ingressInformer.Lister().Ingresses(sc.namespace).List(sc.labelSelector)
if err != nil {
return nil, err
}
ingresses, err = sc.filterByAnnotations(ingresses)
if err != nil {
return nil, err
}
ingresses, err = sc.filterByIngressClass(ingresses)
if err != nil {
return nil, err
}
endpoints := []*endpoint.Endpoint{}
for _, ing := range ingresses {
// Check the controller annotation to see if we are responsible.
if controller, ok := ing.Annotations[annotations.ControllerKey]; ok && controller != annotations.ControllerValue {
log.Debugf("Skipping ingress %s/%s because controller value does not match, found: %s, required: %s",
ing.Namespace, ing.Name, controller, annotations.ControllerValue)
continue
}
ingEndpoints := endpointsFromIngress(ing, sc.ignoreHostnameAnnotation, sc.ignoreIngressTLSSpec, sc.ignoreIngressRulesSpec)
// apply template if host is missing on ingress
if (sc.combineFQDNAnnotation || len(ingEndpoints) == 0) && sc.fqdnTemplate != nil {
iEndpoints, err := sc.endpointsFromTemplate(ing)
if err != nil {
return nil, err
}
ingEndpoints = append(ingEndpoints, iEndpoints...)
}
if len(ingEndpoints) == 0 {
log.Debugf("No endpoints could be generated from ingress %s/%s", ing.Namespace, ing.Name)
continue
}
log.Debugf("Endpoints generated from ingress: %s/%s: %v", ing.Namespace, ing.Name, ingEndpoints)
endpoints = append(endpoints, ingEndpoints...)
}
for _, ep := range endpoints {
sort.Sort(ep.Targets)
}
return endpoints, nil
}
func (sc *ingressSource) endpointsFromTemplate(ing *networkv1.Ingress) ([]*endpoint.Endpoint, error) {
hostnames, err := fqdn.ExecTemplate(sc.fqdnTemplate, ing)
if err != nil {
return nil, err
}
resource := fmt.Sprintf("ingress/%s/%s", ing.Namespace, ing.Name)
ttl := annotations.TTLFromAnnotations(ing.Annotations, resource)
targets := annotations.TargetsFromTargetAnnotation(ing.Annotations)
if len(targets) == 0 {
targets = targetsFromIngressStatus(ing.Status)
}
providerSpecific, setIdentifier := annotations.ProviderSpecificAnnotations(ing.Annotations)
var endpoints []*endpoint.Endpoint
for _, hostname := range hostnames {
endpoints = append(endpoints, EndpointsForHostname(hostname, targets, ttl, providerSpecific, setIdentifier, resource)...)
}
return endpoints, nil
}
// filterByAnnotations filters a list of ingresses by a given annotation selector.
func (sc *ingressSource) filterByAnnotations(ingresses []*networkv1.Ingress) ([]*networkv1.Ingress, error) {
selector, err := getLabelSelector(sc.annotationFilter)
if err != nil {
return nil, err
}
// empty filter returns original list
if selector.Empty() {
return ingresses, nil
}
filteredList := []*networkv1.Ingress{}
for _, ingress := range ingresses {
// include ingress if its annotations match the selector
if matchLabelSelector(selector, ingress.Annotations) {
filteredList = append(filteredList, ingress)
}
}
return filteredList, nil
}
// filterByIngressClass filters a list of ingresses based on a required ingress
// class
func (sc *ingressSource) filterByIngressClass(ingresses []*networkv1.Ingress) ([]*networkv1.Ingress, error) {
// if no class filter is specified then there's nothing to do
if len(sc.ingressClassNames) == 0 {
return ingresses, nil
}
classNameReq, err := labels.NewRequirement(IngressClassAnnotationKey, selection.In, sc.ingressClassNames)
if err != nil {
return nil, err
}
selector := labels.NewSelector()
selector = selector.Add(*classNameReq)
filteredList := []*networkv1.Ingress{}
for _, ingress := range ingresses {
var matched = false
for _, nameFilter := range sc.ingressClassNames {
if ingress.Spec.IngressClassName != nil && len(*ingress.Spec.IngressClassName) > 0 {
if nameFilter == *ingress.Spec.IngressClassName {
matched = true
}
} else if matchLabelSelector(selector, ingress.Annotations) {
matched = true
}
if matched {
filteredList = append(filteredList, ingress)
break
}
}
if !matched {
log.Debugf("Discarding ingress %s/%s because it does not match required ingress classes %v", ingress.Namespace, ingress.Name, sc.ingressClassNames)
}
}
return filteredList, nil
}
// endpointsFromIngress extracts the endpoints from ingress object
func endpointsFromIngress(ing *networkv1.Ingress, ignoreHostnameAnnotation bool, ignoreIngressTLSSpec bool, ignoreIngressRulesSpec bool) []*endpoint.Endpoint {
resource := fmt.Sprintf("ingress/%s/%s", ing.Namespace, ing.Name)
ttl := annotations.TTLFromAnnotations(ing.Annotations, resource)
targets := annotations.TargetsFromTargetAnnotation(ing.Annotations)
if len(targets) == 0 {
targets = targetsFromIngressStatus(ing.Status)
}
providerSpecific, setIdentifier := annotations.ProviderSpecificAnnotations(ing.Annotations)
// Gather endpoints defined on hosts sections of the ingress
var definedHostsEndpoints []*endpoint.Endpoint
// Skip endpoints if we do not want entries from Rules section
if !ignoreIngressRulesSpec {
for _, rule := range ing.Spec.Rules {
if rule.Host == "" {
continue
}
definedHostsEndpoints = append(definedHostsEndpoints, EndpointsForHostname(rule.Host, targets, ttl, providerSpecific, setIdentifier, resource)...)
}
}
// Skip endpoints if we do not want entries from tls spec section
if !ignoreIngressTLSSpec {
for _, tls := range ing.Spec.TLS {
for _, host := range tls.Hosts {
if host == "" {
continue
}
definedHostsEndpoints = append(definedHostsEndpoints, EndpointsForHostname(host, targets, ttl, providerSpecific, setIdentifier, resource)...)
}
}
}
// Gather endpoints defined on annotations in the ingress
var annotationEndpoints []*endpoint.Endpoint
if !ignoreHostnameAnnotation {
for _, hostname := range annotations.HostnamesFromAnnotations(ing.Annotations) {
annotationEndpoints = append(annotationEndpoints, EndpointsForHostname(hostname, targets, ttl, providerSpecific, setIdentifier, resource)...)
}
}
// Determine which hostnames to consider in our final list
hostnameSourceAnnotation, hostnameSourceAnnotationExists := ing.Annotations[annotations.IngressHostnameSourceKey]
if !hostnameSourceAnnotationExists {
return append(definedHostsEndpoints, annotationEndpoints...)
}
// Include endpoints according to the hostname source annotation in our final list
var endpoints []*endpoint.Endpoint
if strings.ToLower(hostnameSourceAnnotation) == IngressHostnameSourceDefinedHostsOnlyValue {
endpoints = append(endpoints, definedHostsEndpoints...)
}
if strings.ToLower(hostnameSourceAnnotation) == IngressHostnameSourceAnnotationOnlyValue {
endpoints = append(endpoints, annotationEndpoints...)
}
return endpoints
}
func targetsFromIngressStatus(status networkv1.IngressStatus) endpoint.Targets {
var targets endpoint.Targets
for _, lb := range status.LoadBalancer.Ingress {
if lb.IP != "" {
targets = append(targets, lb.IP)
}
if lb.Hostname != "" {
targets = append(targets, lb.Hostname)
}
}
return targets
}
func (sc *ingressSource) AddEventHandler(ctx context.Context, handler func()) {
log.Debug("Adding event handler for ingress")
// Right now there is no way to remove event handler from informer, see:
// https://github.com/kubernetes/kubernetes/issues/79610
_, _ = sc.ingressInformer.Informer().AddEventHandler(eventHandlerFunc(handler))
}