/ 
Kubernetes-Comb-0.001
/ Kubernetes::Comb::Upstream::K8s

NAME

Kubernetes::Comb::Upstream::K8s - Borrow a Comb's service from its peer Comb in another kube context

VERSION

version 0.001

SYNOPSIS

# the controlling code: every Comb borrows from its peer in dev
my $comb = Kubernetes::Comb->from_crd($cr,
  upstream => sub { K8s => ( context => 'dev' ) }
);

# the same in the custom resource
upstream:
  class: Kubernetes::Comb::Upstream::K8s
  context: dev
  namespace: platform     # default: the Comb's own
  name: nats              # default: the Comb's own

DESCRIPTION

The layering upstream: the service comes from the peer Comb -- the same Comb, one layer up -- in another kube context. It reads the peer's Comb custom resource and nothing else, so get on combs in that namespace is all the access it needs. The custom resource only names the context; the credentials are those of the local kubeconfig.

The client for the context comes from the Comb's own ("for_context" in Kubernetes::Comb::Role::Client), synchronous or asynchronous alike. The peer is read as the Comb's "crd_class" in Kubernetes::Comb.

What cannot be read -- a context that is missing or broken, a custom resource that is not there or not accessible -- is no error: "status" reports it unreachable, and the Comb goes Blocked with the reason.

Because every Comb publishes its already resolved endpoints in status.endpoints, the peer's addresses are the right ones however many layers are behind it, and "status" reports them in via.

context

Required. The kube context the peer lives in, from the kubeconfig the Comb's own client uses.

namespace

Namespace of the peer. Default: the Comb's own.

name

Name of the peer. Default: the Comb's own.

endpoint_class

The class "endpoints" builds, Kubernetes::Comb::Endpoint.

status

my $seen = $upstream->status($comb)->get;

Future of { reachable, phase, via, context, message }: phase is the peer's, via this "context" followed by the peer's own via. Never fails: whatever keeps the peer from being read makes it unreachable, with the reason in message. So does a peer that is the Comb itself. A message also names the peer's endpoints that have no address reachable from here.

endpoints

my $endpoints = $upstream->endpoints($comb)->get;

Future of the arrayref of the peer's published endpoints that have an address reachable from here, as Kubernetes::Comb::Endpoint. When both contexts point at the same API server ("server_url" in Kubernetes::Comb::Role::Client) cluster is the peer's cluster address, else its external one; external stays the peer's. Fails with the reason when the peer cannot be read.

SEE ALSO

SUPPORT

Issues

Please report bugs and feature requests on GitHub at https://github.com/Getty/p5-kubernetes-comb/issues.

IRC

Join #kubernetes on irc.perl.org or message Getty directly.

CONTRIBUTING

Contributions are welcome! Please fork the repository and submit a pull request.

AUTHOR

Torsten Raudssus <getty@cpan.org>

COPYRIGHT AND LICENSE

This software is copyright (c) 2026 by Torsten Raudssus <torsten@raudssus.de> https://raudssus.de/.

This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.