chore: sync prod into staging to unblock release PR #1346

[suisuss]

Problem

Release PR #1346 (staging -> prod) is DIRTY / CONFLICTING. Single conflict:

• drizzle/meta/_journal.json

Root cause

PR #1340 (release: To Prod) was squash-merged instead of merged as a merge commit. Every previous staging -> prod release was a merge commit (2 parents); #1340 has 1 parent.

sha       parents  subject
28f8c91a  2        Merge pull request #1337 from KeeperHub/staging
3a56c5ce  2        Merge pull request #1333 from KeeperHub/staging
0f218a77  2        Merge pull request #1329 from KeeperHub/staging
14a61bcc  2        Merge pull request #1314 from KeeperHub/staging
aebbed7d  2        Merge pull request #1311 from KeeperHub/staging
10ed1850  2        Merge pull request #1304 from KeeperHub/staging
adaabe1f  1        release: To Prod (drop DB integration block migration) (#1340)   <- squash

A merge commit preserves shared ancestry. A squash takes staging's tree and creates a brand-new commit on prod with no parent link back. So after #1340:

• staging history: hotfix as commit 75eb2849 (from squash-merging hotfix(security): drop global Database Query plugin block #1339 into staging)
• prod history: hotfix as commit adaabe1f (squash of staging into prod)
• the two commits contain byte-identical migration content but are unrelated objects
• merge-base of prod and staging stays at 90f2d7ae -- from before either side had the hotfix

When this PR's parent #1346 opens, git does a 3-way merge from that pre-hotfix base. Both sides modified the tail of _journal.json. Git can't tell the edits are logically the same; it just sees overlapping changes and conflicts.

No one pushed directly to prod. No force push, no branch protection bypass. The wrong option was selected in the GitHub merge dropdown on #1340.

Three-way state on the conflicting file:

Ref	Last journal entry	EOF newline
merge-base 90f2d7ae	idx 82 (block_executions_for_deactivated_users)	no
origin/prod (after #1340)	idx 83 (drop_db_integration_block)	no
origin/staging	idx 84 (employee_only_signup_block)	yes

Resolution

This PR merges origin/prod into staging and keeps staging's version of _journal.json (strict superset: 82 + 83 + 84). The merge restores shared ancestry so PR #1346 becomes a clean merge.

• No migration file content changed.
• git diff origin/staging HEAD is empty for non-merge content.
• git merge-base --is-ancestor origin/prod HEAD is true after this merge.

Once this lands, PR #1346 should auto-flip to MERGEABLE.

Prevention (not in this PR -- needs follow-up)

The repo currently allows all three merge methods on every base branch:

allow_merge_commit:  true
allow_squash_merge:  true   <- caused this
allow_rebase_merge:  true

Options to stop a recurrence:

1. Disable allow_squash_merge repo-wide. Forces merge commits everywhere; feature PRs lose the squash option.
2. CI guard on prod: fail if the new merge commit has fewer than 2 parents. Cheap, targeted, doesn't break feature PR ergonomics. Recommended.
3. Process only -- code-review the merge action. Cheapest, but the same misclick will happen again.

Out of scope

build-images / build on #1346 failed with No space left on device on a self-hosted runner. Infra issue, unrelated to this conflict. Re-run after this PR lands and the runner host is cleaned up.

[github-actions]

🧹 PR Environment Cleaned Up

The PR environment has been successfully deleted.

Deleted Resources:

• Namespace: pr-1347
• All Helm releases (Keeperhub, Scheduler, Event services)
• PostgreSQL Database (including data)
• LocalStack, Redis
• All associated secrets and configs

All resources have been cleaned up and will no longer incur costs.

[github-actions]

ℹ️ No PR Environment to Clean Up

No PR environment was found for this PR. This is expected if:

• The PR never had the deploy-pr-environment label
• The environment was already cleaned up
• The deployment never completed successfully