Title: Feature Request: Support Longer Execution Limits for Serverless Functions and Fluid Compute to Accommodate AI Agent Frameworks (e.g., Eino)
Issue Description: Currently, Vercel Serverless Functions default to a 15-second timeout, which can be configured up to 300 seconds on the Pro plan (and up to 800 seconds with Fluid Compute). However, when deploying AI Agent frameworks like Eino, agents often involve multi-step reasoning, multiple tool-calling loops, or long-form text generation. In these complex scenarios, the execution time easily exceeds current limits, leading to timeouts and limiting the feasibility of running robust agentic workflows on Vercel.
Expected Solution: We hope Vercel can provide longer execution limits specifically optimized for AI Agent workloads, or offer more seamless support for long-running, asynchronous background task execution (such as enhanced support for background coroutines), to fully accommodate the lifecycle of agent frameworks like Eino.
标题： 建议延长 Serverless 函数与 Fluid Compute 的最长运行时间以支持 AI Agent 框架（如 Eino）的部署
问题背景： 目前 Vercel Serverless 函数的默认超时为 15 秒，即使在 Pro 计划中配置 maxDuration 也最多支持 300 秒（Fluid Compute 支持 800 秒）。然而，当部署 Eino 等 AI Agent 开发框架时，Agent 通常需要进行多步骤推理、多次工具调用（Tool Calling）或生成长文本。在这种复杂的长任务场景中，函数极易触发超时错误，这限制了在 Vercel 上运行复杂智能体系统的可行性。
期望的解决方案： 期望 Vercel 能够针对 AI 负载提供更长的函数运行时间配额，或者提供更完善的异步后台长任务执行支持（例如更强大的后台协程处理能力），以便完美适配 Eino 等 Agent 框架的长任务生命周期。