Taskcluster

GENERAL

▶ [patch] [6235](https://github.com/taskcluster/taskcluster/issues/6235)
Migrates to aws sdk v3.

USERS

▶ [patch] [6563](https://github.com/taskcluster/taskcluster/issues/6563)
Ensure livelog and interactive tasks tunnels work after a websocktunnel restart.

USERS

▶ [patch] [6775](https://github.com/taskcluster/taskcluster/issues/6775)
Fixed error messages not being displayed on Pending/Claimed tasks pages in case scopes were missing.

WORKER-DEPLOYERS

▶ [patch] [6779](https://github.com/taskcluster/taskcluster/issues/6779)
Interactive feature data race fixed, whereby an error could cause a concurrent read and write of process state in different go routines.

USERS

▶ [minor] [bug 1874568](http://bugzil.la/1874568)
Fixes token expiry issue for Auth0 login strategy that was not refetched.

▶ [patch] [6495](https://github.com/taskcluster/taskcluster/issues/6495)
Improve changelog rendering in UI

DEVELOPERS

▶ [patch]
Added helper script to make it easier to run websocktunnel locally.

Automated Package Updates

<details>
<summary>1 Dependabot updates</summary>

* build(deps): bump jinja2 from 3.1.2 to 3.1.3 in /taskcluster (c97cf83aa)

</details>

GENERAL

▶ [patch] [6762](https://github.com/taskcluster/taskcluster/issues/6762)
Generic Worker: Tasks internally translated by D2G will add exit code 128 to the retry exit status array for retrying on an intermittent docker image pull issue.

Automated Package Updates

<details>
<summary>5 Dependabot updates</summary>

* build(deps): bump follow-redirects from 1.14.8 to 1.15.4 in /ui (d0f85465b)
* build(deps): bump follow-redirects in /clients/client (fdf26413e)
* build(deps): bump follow-redirects from 1.15.3 to 1.15.4 (77c73c527)
* build(deps): bump follow-redirects in /clients/client-test (1465e064d)
* build(deps): bump follow-redirects in /clients/client-web (924ecbb91)

</details>

GENERAL

▶ [patch]
Generic Worker: No longer logs out the redirect URL for the interactive shell feature since they are signed.

DEPLOYERS

▶ [MAJOR] [6759](https://github.com/taskcluster/taskcluster/issues/6759)
Kubernetes pods use liveness probe instead of unconditionally killing containers daily.

▶ [patch] [6641](https://github.com/taskcluster/taskcluster/issues/6641)
Worker-manager no longer counts "stopping" instances as part of the existing capacity when estimating the number of workers to start (although they are still counted towards maxCapacity).

Automated Package Updates

<details>
<summary>7 Dependabot updates</summary>

* build(deps-dev): bump markdown-table from 2.0.0 to 3.0.3 (979bdebbe)
* build(deps): bump the deps group in /ui with 5 updates (d1b5aedaf)
* build(deps): bump taskcluster from 58.0.2 to 59.2.0 in /taskcluster (dcb3409ae)
* build(deps): bump the deps group with 7 updates (3381ad533)
* build(deps): bump the deps group with 1 update (497c11ff6)
* build(deps): bump the deps group in /taskcluster with 1 update (156da5e47)
* build(deps): bump github/codeql-action from 2 to 3 (1d9788a67)

</details>