feat(source/fake): extend fake source to emit all supported record types by ivankatliarchuk · Pull Request #6308 · kubernetes-sigs/external-dns

ivankatliarchuk · 2026-03-24T08:53:55Z

What does it do ?

source/fake.go — generates one endpoint per type in KnownRecordTypes (A, AAAA, CNAME, TXT, SRV, NS, PTR, MX, NAPTR); adds ResourceLabelKey label mirroring service source; honours --fqdn-template via Config.TemplateEngine; caps fake Pod name at 253 chars
pkg/events/types.go - sets event.Regarding from Name (not gated on UID), so fake source endpoints produce valid Kubernetes events; Related still requires UID
docs/sources/fake.md - new page covering use cases, generated endpoints table, --managed-record-types, --emit-events, and --fqdn-template
100% coverage

Motivation

Webhook providers (3rd-party DNS backends) had no way to test all supported record types without a live Kubernetes cluster. The fake source only emitted A records, making it impossible to verify AAAA, SRV, MX, NAPTR, etc. handling. Additionally, Kubernetes events were silently skipped for fake endpoints because Regarding was never populated (UID gated), and the domain was hardcoded to example.com.

More

Yes, this PR title follows Conventional Commits
Yes, I added unit tests
Yes, I updated end user documentation accordingly

Signed-off-by: ivan katliarchuk <ivan.katliarchuk@gmail.com>

k8s-ci-robot · 2026-03-24T08:54:02Z

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by:
Once this PR has been reviewed and has the lgtm label, please assign szuecs for approval. For more information see the Code Review Process.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

coveralls · 2026-03-24T08:59:39Z

Pull Request Test Coverage Report for Build 23734458438

Details

0 of 0 changed or added relevant lines in 0 files are covered.
4 unchanged lines in 1 file lost coverage.
Overall coverage increased (+0.04%) to 80.515%

Files with Coverage Reduction	New Missed Lines	%
fake.go	4	95.96%

Totals
Change from base Build 23734434195:	0.04%
Covered Lines:	16987
Relevant Lines:	21098

💛 - Coveralls

Signed-off-by: ivan katliarchuk <ivan.katliarchuk@gmail.com>

ivankatliarchuk · 2026-03-24T09:18:21Z

Hi @frittentheke . Have a look, I'm not sure where all records supported, but yeah, make sense to use fake source as some sort of contract/issue tracker.

frittentheke · 2026-03-24T09:23:25Z

Hi @frittentheke . Have a look, I'm not sure where all records supported, but yeah, make sense to use fake source as some sort of contract/issue tracker.

Awesome thank you @ivankatliarchuk .
I started a little on a PR adding IPv6/AAAA and MX ...
But fake providing the more complex records is what helps.

I'll push my code as draft later for you to take ans maybe make use of ....

Edit:

I see you already have lots of code including the use of managed-record-types which I added as well.

Cool 😎

Signed-off-by: ivan katliarchuk <ivan.katliarchuk@gmail.com>

ivankatliarchuk · 2026-03-26T00:06:46Z

Lgtm is enough. We could incorporate other code changes in follow up PRs ;-)

frittentheke

LGTM @ivankatliarchuk

... just dropped a few thoughts on making the test-data even more "valuable" or "challenging" for the consumer.

frittentheke · 2026-03-26T16:08:24Z

+	case endpoint.RecordTypePTR:
+		name := generateDNSName(4, sc.dnsName)
+		var err error
+		ep, err = endpoint.NewPTREndpoint(generateIPv4Address(), endpoint.TTL(0), name)


Maybe randomly generate IPv6 PTR records (ip6.arpa) as well?

I'm not sure. We going to have a flag for that #6232. But in reality it should be done with DNSEndpoint

frittentheke · 2026-03-26T16:12:44Z

+	case endpoint.RecordTypeSRV:
+		// SRV target format: "priority weight port target." (target must end with a dot per RFC 2782)
+		name := generateDNSName(4, sc.dnsName)
+		ep = endpoint.NewEndpoint(fmt.Sprintf("_sip._udp.%s", sc.dnsName), endpoint.RecordTypeSRV, fmt.Sprintf("10 20 5060 %s.", name))


Since SRV records are quite complex, maybe a bit more variety makes sense?
Just to ensure that

I have something like

case endpoint.RecordTypeSRV: // SRV syntax: // SPEC: _service._proto.name. ttl IN SRV priority weight port target. // EXAMPLE: _sip._tcp.example.com. 86400 IN SRV 0 5 5060 sipserver.example.com. var priority int = rand.Intn(100) + 1 var weight int = rand.Intn(100) + 1 var target string = generateDNSName(6, sc.dnsName) + "." var service string = "_" + generateDNSName(5, "") var protocol []string = []string{"_tcp", "_udp"} serviceName = fmt.Sprintf("%s%s ", service, protocol[rand.Intn(2)]) address = fmt.Sprintf("%d %d %s", priority, weight, target)

in mind.

Not sure. Better to use DNSEndpoint, fake source is to verity handling of these types.

frittentheke · 2026-03-26T16:17:36Z

 }

-// Endpoints returns endpoint objects.
+// Endpoints returns one endpoint per supported DNS record type.


Maybe sticking with creating 10 each makes sense?
Since Fake is used for testing webhooks or other integrations have a "bigger" set might be helpful to cover more ground?

I've change it sligthtly. Updated docs, high level - user could have FQDN template to increase number of endpoints

frittentheke · 2026-03-26T17:05:17Z

+		ep.WithLabel(endpoint.ResourceLabelKey, fmt.Sprintf("%s/%s/%s", types.Fake, pod.Namespace, ep.DNSName))
+		ep.WithRefObject(events.NewObjectReference(&pod, types.Fake))
+	}
+	return ep, nil


Maybe add some (debug) logging the Record generated?

frittentheke · 2026-03-26T17:13:14Z

+	var ep *endpoint.Endpoint
+
+	switch recordType {
+	case endpoint.RecordTypeA:


Maybe adding a record containing more than one IP address makes sense?

frittentheke · 2026-03-26T17:13:54Z

+	switch recordType {
+	case endpoint.RecordTypeA:
+		ep = endpoint.NewEndpoint(generateDNSName(4, sc.dnsName), endpoint.RecordTypeA, generateIPv4Address())
+	case endpoint.RecordTypeAAAA:


Same as with IPv4 having a record contain more than one address might make for sensible test data?

ivankatliarchuk · 2026-03-26T19:03:43Z

LGTM @ivankatliarchuk

... just dropped a few thoughts on making the test-data even more "valuable" or "challenging" for the consumer.

I'm a bit on the edge to overload Fake source. Updated docs to better frame the fake source's scope.

The fake source is a smoke-test tool - it answers "does the provider connect and accept records?" without needing a cluster. It is not a substitute for a contract test. For predictable, stable integration test fixtures (specific record types, exact counts, known values), the right approach is DNSEndpoint resources with --source=crd, where the user has full control.

A few things worth noting:

SRV, PTR, and NAPTR are emitted by the fake source, but in practice they are only reachable via the CRD source. The fake source lets webhook providers verify handling of those types, but a passing result doesn't mean any real source will produce them.
The --fqdn-template comma-separated list adds volume (useful for load/stress testing) but not gurantee predictability

Documentation updated accordingly.

  ┌───────────────────┬──────────────────────────────────────────────────────────────┬───────────────────────────────────────────────────────────────┐
  │                   │                         Fake source                          │                        DNSEndpoint CRD                        │
  ├───────────────────┼──────────────────────────────────────────────────────────────┼───────────────────────────────────────────────────────────────┤
  │ Purpose           │ Smoke test — "does the provider connect and accept records?" │ Integration test — predictable, exact record types and counts │
  ├───────────────────┼──────────────────────────────────────────────────────────────┼───────────────────────────────────────────────────────────────┤
  │ Record types      │ Whatever fake emits                                          │ You decide                                                    │
  ├───────────────────┼──────────────────────────────────────────────────────────────┼───────────────────────────────────────────────────────────────┤
  │ Volume            │ FQDN template hack                                           │ You decide                                                    │
  ├───────────────────┼──────────────────────────────────────────────────────────────┼───────────────────────────────────────────────────────────────┤
  │ PTR / SRV / NAPTR │ Fake emits them, but real sources only CRD                  │ Explicitly supported                                          │
  ├───────────────────┼──────────────────────────────────────────────────────────────┼───────────────────────────────────────────────────────────────┤
  │ Predictability    │ Random names, random IPs, changes every reconciliation       │ Stable, declarative                                           │
  └───────────────────┴──────────────────────────────────────────────────────────────┴───────────────────────────────────────────────────────────────┘

Signed-off-by: ivan katliarchuk <ivan.katliarchuk@gmail.com>

…d-types

Signed-off-by: ivan katliarchuk <ivan.katliarchuk@gmail.com>

frittentheke · 2026-03-27T09:51:01Z

Thanks @ivankatliarchuk ! You are right, this PR is big enough, but it's also more than a leap forward.
Yes, before doing even more, the question about the actual role and intention of the Fake source should be discussed.

In the end it's not about creating and maintaining all possible values for all sorts of records and to validate a consumer does accept them. But I thank you again for picking up on the idea of providing just "a little" more test data for webhook authors to ensure their part works and to allow for some lightweight CI testing.

ivankatliarchuk · 2026-03-29T11:49:24Z

We don't hold any sessions where we discuss things ;-). I've added a doc, to clarify them

…d-types

feat(source/fake): extend fake source to emit all supported record types

cdadf49

Signed-off-by: ivan katliarchuk <ivan.katliarchuk@gmail.com>

k8s-ci-robot requested review from mloiseleur and vflaux March 24, 2026 08:54

k8s-ci-robot added docs source size/L Denotes a PR that changes 100-499 lines, ignoring generated files. cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. labels Mar 24, 2026

feat(source/fake): extend fake source to emit all supported record types

681f729

Signed-off-by: ivan katliarchuk <ivan.katliarchuk@gmail.com>

ivankatliarchuk added 2 commits March 24, 2026 09:28

feat(source/fake): extend fake source to emit all supported record types

60cd8a4

Signed-off-by: ivan katliarchuk <ivan.katliarchuk@gmail.com>

feat(source/fake): extend fake source to emit all supported record types

9d21f25

Signed-off-by: ivan katliarchuk <ivan.katliarchuk@gmail.com>

frittentheke reviewed Mar 26, 2026

View reviewed changes

ivankatliarchuk and others added 2 commits March 26, 2026 19:03

feat(source/fake): extend fake source to emit all supported record types

c1167b0

Signed-off-by: ivan katliarchuk <ivan.katliarchuk@gmail.com>

Merge branch 'kubernetes-sigs:master' into feat/fake-source-all-recor…

2c92473

…d-types

k8s-ci-robot added size/XL Denotes a PR that changes 500-999 lines, ignoring generated files. and removed size/L Denotes a PR that changes 100-499 lines, ignoring generated files. labels Mar 26, 2026

feat(source/fake): extend fake source to emit all supported record types

46994d1

Signed-off-by: ivan katliarchuk <ivan.katliarchuk@gmail.com>

ivankatliarchuk added 2 commits March 29, 2026 22:11

Merge branch 'kubernetes-sigs:master' into feat/fake-source-all-recor…

6e694e8

…d-types

Merge branch 'kubernetes-sigs:master' into feat/fake-source-all-recor…

18d9a73

…d-types

frittentheke mentioned this pull request Apr 3, 2026

Add support for NS delegation and MX records inovex/external-dns-openstack-webhook#70

Open

Conversation

ivankatliarchuk commented Mar 24, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What does it do ?

Motivation

More

Uh oh!

k8s-ci-robot commented Mar 24, 2026

Uh oh!

coveralls commented Mar 24, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Pull Request Test Coverage Report for Build 23734458438

Details

💛 - Coveralls

Uh oh!

ivankatliarchuk commented Mar 24, 2026

Uh oh!

frittentheke commented Mar 24, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ivankatliarchuk commented Mar 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

frittentheke left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ivankatliarchuk commented Mar 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

frittentheke commented Mar 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ivankatliarchuk commented Mar 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

ivankatliarchuk commented Mar 24, 2026 •

edited

Loading

coveralls commented Mar 24, 2026 •

edited

Loading

frittentheke commented Mar 24, 2026 •

edited

Loading

ivankatliarchuk commented Mar 26, 2026 •

edited

Loading

frittentheke left a comment •

edited

Loading

ivankatliarchuk commented Mar 26, 2026 •

edited

Loading

frittentheke commented Mar 27, 2026 •

edited

Loading

ivankatliarchuk commented Mar 29, 2026 •

edited

Loading