LB health checks failing on v1.5.1

[m1pl]

I deployed a kubeflow cluster following the docs with the v1.5.1 tag. Everything was successful, but instead of the dashboard I see

Error: Server Error
The server encountered a temporary error and could not complete your request.
Please try again in 30 seconds.

The LB in GCE shows this:

The exact same deployment with tag v1.5.0 is working fine.

Any help would be appreciated.

[skrawn]

I ran into the same issue, you need to patch two files to fix the health check for backend service: #371

[gkcalat]

This is the same as #371. It was fixed in #374.

You can pull the latest version from master branch and follow the deployment instructions or wait for Kubeflow v1.6 release, which is expected in August 2022.