You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@flink.apache.org by Daniel Rossos via user <us...@flink.apache.org> on 2025/10/27 18:18:10 UTC
[DISCUSSION] FlinkBlueGreenDeployment Kubernetes Networking
Hello everyone,
Starting a discussion thread about improvements that can be made to the
Kubernetes networking attached to the FlinkBlueGreenDeployments. I have
raised the issue that prompted this discussion in FLINK-38577
<https://issues.apache.org/jira/browse/FLINK-38577> and wanted to see what
approaches people are thinking would be best.
As mentioned in the ticket the first thoughts I had would be something like
1 svc and 1 ingress that will dynamically bind to whatever FlinkDeployment
is active. That way would avoid the need for multiple services and
ingresses to keep coming up and keep naming + address of those resources
stable despite the underlying Flink deployments changing.
Have yet to prototype that though (will try to get something early this
week there), but would love preliminary feedback and thoughts.
Thanks,
Daniel
Re: [DISCUSSION] FlinkBlueGreenDeployment Kubernetes Networking
Posted by Daniel Rossos via user <us...@flink.apache.org>.
Oops, sent this discussion to wrong mailing list, please disregard meant to
go to dev
On Mon, Oct 27, 2025 at 2:18 PM Daniel Rossos <da...@shopify.com>
wrote:
> Hello everyone,
>
> Starting a discussion thread about improvements that can be made to the
> Kubernetes networking attached to the FlinkBlueGreenDeployments. I have
> raised the issue that prompted this discussion in FLINK-38577
> <https://issues.apache.org/jira/browse/FLINK-38577> and wanted to see
> what approaches people are thinking would be best.
>
> As mentioned in the ticket the first thoughts I had would be something
> like 1 svc and 1 ingress that will dynamically bind to whatever
> FlinkDeployment is active. That way would avoid the need for multiple
> services and ingresses to keep coming up and keep naming + address of those
> resources stable despite the underlying Flink deployments changing.
>
> Have yet to prototype that though (will try to get something early this
> week there), but would love preliminary feedback and thoughts.
>
> Thanks,
>
> Daniel
>