The Stop List for Free Inference and Free Hosts
说明与用途
团队创建了一个“停止列表”,将免费依赖与风险信号和预先约定的操作关联起来。通过使用JSON账本、lint检查和分离演练,他们建立了一个可重复的流程,用于决定何时放弃免费推理服务。
问题
Teams adopt free inference endpoints or servers without clear criteria for when to abandon them, leading to hidden risks beyond cost.
结果
Teams gain a repeatable, low‑overhead process to evaluate and exit free inference dependencies, reducing hidden operational risk.
工作流
- Define a capability ledger JSON per job listing endpoint, capabilities, latency, fallback, and owner.
- Lint the ledger with a CI pre‑merge check that flags red‑line capabilities such as "mutates_durable_state".
- Run a detach drill by pointing provider base URLs to a port that refuses connections and compare job behavior to the normal run.
- If the job fails or behaves unexpectedly, decide to add a second writer or mark the job as an accepted exception.
- Maintain an exit‑criteria table to record agreed actions for each risk signal.
- Use the stop list to allow on‑call engineers to execute the exit decision at any time without committee.
证据
A stop list fixes this. It is a short document, dated, that pairs each free dependency with an observable signal and a pre‑agreed action.
Running it against the sample above exits 1 with one line naming the durable write.
Exit criteria are worthless if nobody can measure the exit. The drill runs one command twice: once normally, once with every provider base URL pointed at a port that refuses connections.
- 记录日期: 2026-10-17T07:56:05.519Z