fix: retain the production QStash token region

[suguanYang]

Summary

Production webhook publication failed with user not found in this region (eu-central-1) because the QStash SDK used its default region while our existing token belongs to us-east-1. We now retain QSTASH_BASE_URL=https://qstash-us-east-1.upstash.io in both API and worker task definitions on each production release.

The renderer accepts an optional regional URL; other environments keep the SDK default when the override is absent. No job API or database contract changes are required. Hipo #115 contains the receiving application change and sanitized real-call verification evidence.

Verification

• ECS task-definition contracts: 10 passed (pytest deploy/ecs/test_render_task_definitions.py -q -o addopts=).
• Ruff checks passed for both changed Python files; git diff --check passed.
• The approved production repair is already running in knowhere-api-prod:56 and knowhere-worker-prod:52. Two real 40,960-byte Markdown uploads through Hipo staging received successful callbacks without retries, 0.304 s and 0.419 s after Knowhere completion.
• These samples verify Markdown and webhook delivery; PDF, DOCX and OCR timings were not measured.

Deployment Notes

• This persists the approved production QSTASH_BASE_URL environment value so later releases do not overwrite the repair. The public callback base URL remains a separate configuration value.
• Existing token, images and secret references are retained. No new credentials, migrations or queue topology changes.
• Removing the override restores the SDK default and would reintroduce the production token-region mismatch.

Checklist

• Tests were added or updated when behavior changed
• Public docs, examples, or OpenAPI contracts were updated when needed
• Database migrations are idempotent and safe to deploy (none)
• Logs, errors, and validation paths avoid leaking secrets or user data
• The pull request description explains any breaking or user-visible change