Skip to content

Conversation

wt
Copy link

@wt wt commented Jul 22, 2025

This chart is meant to be a starting point. This chart will fully fire up an instance of Rauthy in a k8s cluster. However, you will need some form of smtp since rauthy crashes without it. This chart assumes the use of mailcrab, which is a test smtp server that receives mail and present a them in a web UI for testing.

The chart does not require a custom values yaml to launch and is configured for testing. Any serious deployment would need to persist some secrets that are generated in this chart.

Also, this chart runs rauthy on http inside the cluster. I am not sure of a generic way to setup TLS for communication between the ingress proxy and the services inside the cluster. If there is a good generic solution that works be default with all ingress setups, I think that would be a obvious next step.

This chart is meant to be a starting point. This chart will fully fire
up an instance of Rauthy in a k8s cluster. However, you will need some
form of smtp since rauthy crashes without it. This chart assumes the use
of mailcrab, which is a test smtp server that receives mail and present
a them in a web UI for testing.

The chart does not require a custom values yaml to launch and is
configured for testing. Any serious deployment would need to persist
some secrets that are generated in this chart.

Also, this chart runs rauthy on http inside the cluster. I am not sure
of a generic way to setup TLS for communication between the ingress
proxy and the services inside the cluster. If there is a good generic
solution that works be default with all ingress setups, I think that
would be a obvious next step.
@wt
Copy link
Author

wt commented Jul 22, 2025

References #131

@sebadob
Copy link
Owner

sebadob commented Jul 23, 2025

Looks quite nice for testing, but a lot of stuff is missing to make it actually useful of course.
However, one very important thing is that any HPA must not exist because of the Raft and how it's built and maintained. You can scale instances up (at a cost, and if configured properly), but not scale them down afterwards again. The autoscaling should absolutely be removed at least.

I would also not care about adding default variables like for password hashing, when they are hardcoded anyway.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants