Feature/supabase 2025

[luizfelmach]

Feature: Migrate to Supabase 2025 Updates

This pull request introduces significant updates to the supabase-kubernetes Helm chart to align with the latest changes and features in the Supabase 2025 platform.

The primary goal of this migration is to ensure compatibility, incorporate new best practices, and resolve critical issues related to the deployment of a self-hosted Supabase instance on Kubernetes.

Key Changes

refactor: migrate project to Supabase 2025 updates

This commit includes the bulk of the migration work, which involves updating configurations and scripts to match the Supabase 2025 architecture.

• Helm Chart Version Update: The chart version in charts/supabase/Chart.yaml has been incremented from 0.1.3 to 0.2.0.
• Database Initialization Scripts Refactor: Several SQL initialization scripts have been updated, including changes to how JWT secrets and expiration times are handled in 99-jwt.sql.
• New Service Schemas: Introduction of new schemas and roles, such as _analytics and _supavisor, in the database initialization scripts (97-_supabase.sql, 99-logs.sql, 99-pooler.sql, 99-realtime.sql). This suggests an update to the logging, connection pooling, and real-time components.
• Webhook Function Update: The supabase_functions.http_request() function in 98-webhooks.sql has been modified, likely to update its signature or internal logic to align with the latest pg_net extension usage or Supabase's function calling standards.
• Deployment Manifest Updates: Changes across various deployment files (e.g., charts/supabase/templates/db/statefulset.yaml, charts/supabase/templates/realtime/deployment.yaml, charts/supabase/templates/studio/deployment.yaml) to reflect new image tags, environment variables, and configuration mounts for the updated services.

This PR is ready for review and merge. Please let me know if you have any questions.

[zhongjis]

thank you for the work!
just gave it a try and it worked! if anyone ever interested here are my values for deployment

only complaint i have would be the chart still freaks out when using unencrypted db.password. as in the deployment the chart always expect encrypted version

ill start using through out the week and let you know if there's anything else

[lxup]

Thanks for the update @luizfelmach !

I was planning to update the supabase-community Helm chart as well, so I’m checking before duplicating work.
I see the _supavisor schema in the DB init scripts, but no actual supavisor deployment. Is it intentionally missing, or do you plan to add it?

Also, does this chart fully match the latest official supabase/docker-compose setup?

[luizfelmach]

Thanks for the update @luizfelmach !

I was planning to update the supabase-community Helm chart as well, so I’m checking before duplicating work. I see the _supavisor schema in the DB init scripts, but no actual supavisor deployment. Is it intentionally missing, or do you plan to add it?

Also, does this chart fully match the latest official supabase/docker-compose setup?

It’s important to keep it updated with what exists in https://github.com/supabase/supabase/blob/master/docker/docker-compose.yml. I didn’t deploy Supavisor, but it would be very interesting to include it. I believe the only thing missing is that component.

[aantti]

This is some monumental work here @luizfelmach, appreciate it! :)

[propilideno]

We’ve been using this version in our staging environment for two weeks now, and it looks good.
Thanks @luizfelmach.

[aantti]

As the state of this repo hasn't changed much for some time, I'm very much leaning towards accepting this PR as-is to bring it back to a working condition :) → then getting back to the accumulated issues and PRs and figuring out what was solved and can be closed, and what still requires work. Thoughts, @luizfelmach and everyone?

Cc @arpagon

Offtopic: @arpagon - tried to reach via internal channels; @luizfelmach - sent you an email :)

[luizfelmach]

As the state of this repo hasn't changed much for some time, I'm very much leaning towards accepting this PR as-is to bring it back to a working condition :) → then getting back to the accumulated issues and PRs and figuring out what was solved and can be closed, and what still requires work. Thoughts, @luizfelmach and everyone?

Cc @arpagon

Offtopic: @arpagon - tried to reach via internal channels; @luizfelmach - sent you an email :)

Sounds good to me. Getting this PR in to bring the repo back to a working state and then going through the backlog of issues/PRs to see what’s already addressed vs. what still needs work seems like a very reasonable approach.

I can help review and triage the existing issues and PRs after this is merged.

[aantti]

Sounds great, @luizfelmach - very much appreciated and looking forward.

[luizfelmach]

Apparently the pipeline failed due to formatting issues and tests that weren't updated. The last two commits fix this. @aantti.

[aantti]

I'm approving this as the new baseline. @luizfelmach will review the accumulated issues and PRs. A lot of stuff has been already solved by this PR #131, but there are still quite a few things in the backlog.

Related: #120