Retry or exit when OIDC discovery failed/error

[cywteow]

Description

Why there aren't any retry or exit (so that docker can restart by itself) when user specify OIDC settings when it failed to discover or encountered an error. The problem occurred when we restart our server, there can be a race condition between our services e.g. nginx/auth server (authentik) depending who is up before headplane connect using OIDC. Apps should be designed to be robust so that when it's configured to use OIDC as login, even if there's error or failure there is retry or exit (to be restarted automatically).
The current behavior in 0.6.1 logs and error and continue on. The web remains running and the login is stuck because it can't discover oidc.

Headplane Version

0.6.1

Headscale Version

0.27

[tale]

I could make it retry, I'm pretty sure the way it is currently done is the exact same way that Headscale does it? Why not just have docker services wait for each other, make sure the authentication one is healthy before you need to start Headplane?

[bluewalk]

I could make it retry, I'm pretty sure the way it is currently done is the exact same way that Headscale does it? Why not just have docker services wait for each other, make sure the authentication one is healthy before you need to start Headplane?

Because the OIDC provider might be on another machine which could be either down or rebooting upon the start of headplane :) So a retry would definitely solve lots of issues. Might even think about caching the openid-configuration after discovery as a fallback?