cycles: transfer_cycle_issues non-atomic - progress_snapshot saved but issues not moved if process killed mid-transfer

Bug Description

ransfer_cycle_issues performs two DB writes sequentially without wrapping them in 	ransaction.atomic. If the process is killed or a DB error occurs after the first write completes but before the second, the database is left in an inconsistent state with no recovery path.

Affected file

pps/api/plane/utils/cycle_transfer_issues.py:

`python
current_cycle.save(update_fields=["progress_snapshot"]) # line 432 - committed first

... build updated_cycles list ...

CycleIssue.objects.bulk_update(updated_cycles, ["cycle_id"], batch_size=100) # line 458 - may never run
`

Failure scenario

1. Admin triggers "Transfer incomplete issues" from cycle A to cycle B.
2. current_cycle.save(update_fields=["progress_snapshot"]) commits - cycle A now appears complete in the UI.
3. Worker process is killed (OOM kill, deploy restart, DB connection timeout) before CycleIssue.objects.bulk_update runs.
4. Cycle A shows as completed with a snapshot, but all its issues still belong to cycle A - they were never moved.
5. Users see a "completed" cycle full of issues with no way to trigger the transfer again through the UI.

Fix

`python
from django.db import transaction

with transaction.atomic():
current_cycle.save(update_fields=["progress_snapshot"])
CycleIssue.objects.bulk_update(updated_cycles, ["cycle_id"], batch_size=100)
`

Environment

Plane develop branch (2026-08-13).

Start in apps/api/plane/utils/cycle_transfer_issues.py at transfer_cycle_issues and inspect the two writes around lines 432 and 458. Verify that the progress snapshot and CycleIssue updates share one transaction, and confirm that a failure between them does not leave the cycle marked complete while its issues remain unmoved.