fix(ci): harden staging deploy over ssh

[ailuckly]

Add SSH keepalive options to the staging deploy and failure log collection steps so long remote builds do not drop idle connections.

Stop forcing Docker Compose no-cache builds on every deploy; use --pull while preserving build and Maven cache layers to reduce staging deploy time.

Verification: git diff --check; ruby YAML.load_file('.github/workflows/cd-staging.yml')

📌 变更内容

• 如:新增用户登录 API
• 如:修复登录失败时错误提示

✅ 测试验证

• 本地运行通过
• 自测通过
• CI 流水线通过

---

PR 提交规范提醒：

• 确保提交主题信息符合约定式提交规范 (feat/fix/docs/style/refactor/test/chore)
• 确保代码已经通过本地测试
• 确保没有提交敏感信息(密码、密钥等)

[Copilot]

Pull request overview

Hardens the staging CD GitHub Actions workflow to make long-running SSH-based deployments more reliable, and reduces deploy time by avoiding forced no-cache Docker Compose builds.

Changes:

• Added SSH keepalive options (ServerAliveInterval, ServerAliveCountMax, TCPKeepAlive) to both the deploy and failure-log-collection SSH steps.
• Switched docker compose build from --no-cache to --pull to preserve build cache while still refreshing base images for built services.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.