fix(ci): correct stale Helm repository URLs in Release Charts workflow

[Vervious]

Tracking issue

Related to #16

Why are the changes needed?

The Release Charts workflow — the chart-releaser job that publishes to gh-pages, i.e. the only thing that puts charts on https://exa-labs.github.io/flyte — has been failing on every push to master since at least the #14 merge in June. It dies in the "Add required Helm repositories" step, before chart-releaser ever runs, so gh-pages has been frozen at flyte-binary v0.1.10.

Two of the three repository URLs it adds no longer exist:

https://helm.twun.io/                      -> DNS NXDOMAIN
https://kubernetes.github.io/dashboard/    -> 404

Neither is what the charts actually declare. charts/flyte-sandbox/Chart.yaml (and validate-helm-charts.yaml, which is why chart validation still passes) point at the live URLs. This went unnoticed because nothing needed a newly published chart until now — flyte-binary v0.1.11 was merged in #16 and never made it to the index.

What changes were proposed in this pull request?

Point the two stale URLs at the ones the charts declare:

- helm repo add twun https://helm.twun.io/
- helm repo add kubernetes-dashboard https://kubernetes.github.io/dashboard/
+ helm repo add twun https://twuni.github.io/docker-registry.helm
+ helm repo add kubernetes-dashboard https://kubernetes-retired.github.io/dashboard/

Bitnami is left alone — see testing below.

Out of scope: the separate Package & Push Flyte Helm Charts workflow also fails, with Username and password required, because FLYTE_BOT_USERNAME/FLYTE_BOT_PAT aren't set on this fork. That's an upstream GHCR dev-chart path we don't consume, and it's a secrets gap rather than something a diff can fix.

How was this patch tested?

Ran the corrected step's exact command sequence locally, then resolved dependencies for both charts:

$ helm repo add twun https://twuni.github.io/docker-registry.helm
$ helm repo add kubernetes-dashboard https://kubernetes-retired.github.io/dashboard/
$ helm repo add bitnami https://charts.bitnami.com/bitnami
$ helm repo update
$ helm dep update charts/flyte-sandbox
docker-registry       2.2.2     ok
flyte-binary          v0.1.11   ok
kubernetes-dashboard  6.0.0     ok
minio                 12.6.7    ok
postgresql            12.8.1    ok
$ helm dep update charts/flyte-binary
flyteconnector        v0.1.10   ok

helm lint passes on both charts.

Bitnami was checked explicitly rather than assumed, since Bitnami has been retiring its legacy chart index: https://charts.bitnami.com/bitnami 302-redirects to repo.broadcom.com but still serves a valid index.yaml, and both pinned versions (minio 12.6.7, postgresql 12.8.1) download as valid archives. It works today, but it is a redirect away from breaking this workflow the same way.

The real verification is the next push to master: v0.1.11 should appear in https://exa-labs.github.io/flyte/index.yaml.

Labels

fixed

Check all the applicable boxes

• I updated the documentation accordingly. (n/a — CI-only change)
• All new and existing tests passed.
• All commits are signed-off.

Related PRs

• Make flyte-binary safe to run with multiple replicas #16 — the HA change whose flyte-binary v0.1.11 is stuck unpublished because of this.

Link to Devin session: https://app.devin.ai/sessions/137e980b425d43668660198d80b6e21d
Requested by: @Vervious

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment, CI, and merge conflict monitoring