Skip to content

Release 2026-04-20 - (expected chart version 5.30.0)#5194

Merged
fisx merged 31 commits intomasterfrom
release_2026-04-20_15_03
Apr 21, 2026
Merged

Release 2026-04-20 - (expected chart version 5.30.0)#5194
fisx merged 31 commits intomasterfrom
release_2026-04-20_15_03

Conversation

@zebot
Copy link
Copy Markdown
Contributor

@zebot zebot commented Apr 20, 2026

[2026-04-20] (Chart Release 5.30.0)

Release notes

  • Since 5.29 was broken and you should have skipped it, you are about to upgrade from 5.28 to 5.30 in one step. This has been tested and should work. Please consult the release notes from both 5.29 and 5.30 for changes w.r.t. 5.28.

  • background-worker now reuses galley's configmap and secrets for cassandra, postgres and federation domain settings. This removes redundant settings and keeps the two services aligned. No operator action is strictly required; however, we advise removing the background-worker value overrides for galley's cassandra, postgres, and federation domain settings, as they are duplicated and no longer needed:

  • Operators upgrading from the previous wire-server chart release, where the service charts were consolidated into the umbrella chart, must now set tags.proxy explicitly again.

    If your currently installed values no longer contain a proxy tag because of that consolidation, add one before upgrading to this release and set it to the intended state:

  • The Restund helm chart and code stops being supported and shipped. If you have not already, please migrate to coturn which continues to be supported. (Sunset restund (and rex) #5162)

Features

Bug fixes and other updates

Documentation

Internal changes

fisx and others added 30 commits March 24, 2026 18:40
* Changelog.

* Drive-by: cleanup testRetrieveUsersIncludingApps.

* Integration tests.

* Implementation replace app-created event with teams.member-join.

* Drive-by: fix background-worker unit test.
Master->Develop after release
For people using k8ssandra-test-cluster on an existing deployment, the
following settings are needed for an in-cluster upgrade to work by
default.
* Use sbomnix to provide SBOMs for nix generated Docker images

sbomnix creates much better SBOMs than tom-bombadil. Also, we can adhere
to our new rule to have one SBOM file per thing.

* Provide SBOMs for devShells

* Adjust values for wire-server charts inlining

Also, fail if this goes wrong to spot missing template replacements
early.

* Add changelog entry

* Lint
…tween scopes. (#5151)

* Script for generating kotlin and typescript code from swagger.json.

* openapi3 aka swagger: Make schema names unique(r).

With this change, schema names contain a hash of the fully qualified
name of the haskell type.  This should avoid collisions in the hash
table keeping track of all the schema references in openapi3.

* Remove now unused name arguemnts from object, enum.

* Keep named variants of object, enum, ... around for corner cases.

* Versioned schemas.

* Drive-by test coverage extension.
* Test: conversations with apps accross teams.

* Fix: no connections to apps except inside same team.

* Tidy up imports.

* Changelog.

* Reduce number of across-backend requests for connection management.
we want to check substrings the right way round (does the ingressClass contain 'nginx' - function arguments were given the wrong way around)
* contact search does not find apps from outside team.

* "get-app" does not return apps from other teams.

* Extend integration test coverage to federated case.

* Fix unit tests: do not generate arbitrary bots when asked for users.

* Fix integration test: don't allow non-team users to see apps.

* Fix: make getApp verify team as well.

* Fix: fetching profiles from un-connectable apps is fine.

* Remove unused UserProfileFilter constructors and allow app profiles.

App profiles are now shown like regular profiles everywhere except in
end-points "get-app", "get-apps".

* Access control: you can only call "get-apps" on your own team.

* Fix: also list bots as users.

---
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
We're creating SBOMs with different tools now.
* Changelog.

* Integration test.

* Implementation.

- add SetAccountStatus, GetAppIdsForTeam to BrigAPIAccess,
- remove UserProfileFilter,
- revert GetLocalUserProfilesFiltered to GetLocalUserProfiles,
- introduce GetLocalAppProfilesOnly (dedicated & scalable),
- instance SetFeatureConfig AppsConfig now (un-)suspends all apps in team.

* Add note of caution for future me.

* Minimize redundant DB access.

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Better haddocks.

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Revert "Better haddocks."

This reverts commit 82eb5b3.

* Fix imports.

* Drop redundant "only" from effect action name.

* Tweak opportunities (high-hanging, small fruits, though).

* Make zipping less brittle when adding app data to user profile.

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
@zebot zebot requested review from a team as code owners April 20, 2026 15:03
@zebot zebot added the ok-to-test Approved for running tests in CI, overrides not-ok-to-test if both labels exist label Apr 20, 2026
@fisx fisx merged commit 1fa1212 into master Apr 21, 2026
10 checks passed
@fisx fisx deleted the release_2026-04-20_15_03 branch April 21, 2026 13:13
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

ok-to-test Approved for running tests in CI, overrides not-ok-to-test if both labels exist

Projects

None yet

Development

Successfully merging this pull request may close these issues.

8 participants