feat: improve gifts indexing queue priority, task timeouts, and refresh UX feedback

[danoctua]

Description

This PR addresses queue priority ordering for Telegram gift indexing and improves the user experience during manual gift refreshes:

1. Celery Priority Queue Order: Previously, user-initiated high-priority gift indexing tasks could be delayed behind recurring background batches due to Kombu's priority_cycle
   iterating over a Python set with randomized hash seeds in Redis BRPOP. We resolved this by prefixing the queue names alphabetically and using Celery's "queue_order_strategy": "sorted", combined with task_acks_late: True and worker_prefetch_multiplier: 1 for deterministic priority consumption.
2. Backend Task Status Timeout: The /system/async-tasks/{task_id} polling endpoint previously waited indefinitely in a while not task_result.ready() loop. It now enforces a
   default timeout (DEFAULT_ASYNC_TASK_WAIT_TIMEOUT_SECONDS = 9.5) and returns HTTP 408 Request Timeout ("Task took longer than expected") if the task is still executing.
3. Frontend Timeout & UX Feedback: Added targeted per-request timeout support to ApiService (ASYNC_TASK_TIMEOUT_MS = 10000) for task status checks. With Ky's built-in 1-retry
   on status 408, the client waits ~19–20 seconds total before giving up. Added visual feedback toasts in ChatConditions:
   • Success toast on completion ("Gifts refreshed successfully").
   • Warning toast on timeout ("Your request to refresh gifts was submitted, but took longer than expected. Please, wait for 5 minutes or try again"), while refreshing chat state.
   • Error toast on failures or rate limits (e.g. 5-minute cooldown).

Fixes: # (if applicable)

Checklist

Before submitting your pull request, please ensure the following:

• I have read the contributing guidelines.
• I ran all tests and they passed successfully.
• I added/updated documentation (if applicable).
• I reviewed my own code and followed the project's code style.
• I included relevant details in the PR description or commit messages.

Changes

• Queue Prioritization (core/constants.py, indexer_gifts/celery_app.py, docker-compose.yml):
  • Prefixed gift queues with 0-gift-user-priority, 1-gift-collections-sync, and 2-gift-user-batch.
  • Configured "queue_order_strategy": "sorted", "task_acks_late": True, and "worker_prefetch_multiplier": 1.
• Backend Task Polling Timeout (core/constants.py, core/utils/task.py, api/routes/system.py):
  • Added DEFAULT_ASYNC_TASK_WAIT_TIMEOUT_SECONDS = 9.5.
  • Added timeout parameter to wait_for_task raising TimeoutError when exceeded.
  • Renamed route handler to get_task_status, catching TimeoutError and returning HTTP 408 Request Timeout.
• Frontend Targeted Timeout (common/utils/constants.ts, services/ApiService.ts, store/user/api.ts):
  • Defined ASYNC_TASK_TIMEOUT_MS = 10000.
  • Added optional timeout override in ApiOptions and mapped Ky TimeoutError to HTTP 408.
  • Passed timeout override specifically in completeChatTasksAPI.
• Store & UI Feedback (store/user/user.ts, ChatConditions.tsx):
  • Exported TaskTimeoutError for HTTP 408 responses.
  • Added success, warning (timeout), and error feedback toasts in handleRefreshGifts.

How Has This Been Tested?

• Backend Test Suite: Ran make test inside the Docker test environment (all 111 tests passed in 9.97s).
• Backend Linting: Ran pre-commit (ruff, ruff-format) on all modified backend files with zero violations.
• Frontend Build: Verified TypeScript compilation and bundling via npm run build (built cleanly with 0 errors).

Screenshots (if applicable)

N/A

Additional Notes

• The 9.5s backend wait paired with Ky's default 1-retry policy on status 408 provides an optimal 2-attempt window (~19–20s total) allowing tasks that take just slightly longer to
  succeed without blocking the UI indefinitely.