Api's are not using proxy url

[subi-kp]

Summary

I am running cronicle in kubernetes as a docker container. And exposed UI using ingress.

Websocket is working fine but while clicking login button, after entering password i am getting 500.

Error: Network Error: 500: An unknown network error occurred. Please try again.
I inspected from my browser, the request url is hitting internal hostname of the container instead of my endpoint url.
container url: http://cronicle-75987bb6b6-8j7w9:3012/api/user/login --> (current issue )
endpoint url: http://cronicle.xyz.com:3012/api/user/login --> (expected)

[jhuckaby]

This is actually as designed. See: #63 for details.

[subi-kp]

This will work on same network. but in my case container is a private network and meanwhile, I am accessing from public.
Websocket is working perfectly for me, but not apis.

Thanks for your fast response.

[jhuckaby]

That's really bizarre. The websocket and API requests are both supposed to hit the same endpoint. I guess this is an open bug.

[subi-kp]

Can I get any ETA for this bug?

[jhuckaby]

I'm sorry, but I don't do ETAs. This is an open-source free project, that I only work on in my spare time. I also have never used Docker before, so I am not even sure how to reproduce this bug.

You might want to take a look at this pull request: #65

Perhaps his fork corrects this issue.

Apologies.

[subi-kp]

Thanks for your valuable time. Actually I looked into the PR, where the issue is about the web socket port. But in my case url endpoint is an issue, while running cronicle over the nginx proxy. Will send you steps to reproduce it soon

[dol-leodagan]

I'm currently working on a Docker Image for Cronicle, served behind a nginx reverse-proxy and I made a workaround in the entrypoint for this error :

export HOSTNAME="${VIRTUAL_HOST}"

VIRTUAL_HOST is the FQDN of the service

This can bring other problems if you arbitrarily change the HOSTNAME after storage setup but that's not a limitation for my usage...