Recovery

The host watches llama-server. If it crashes or runs out of memory, the host restarts it, at most 3 times per 15 minutes (after 2, 10 and 30 seconds). A server that is only busy or slow to answer /health is never restarted.

What clients see

• A request refused before the model accepted it is retried once after the restart.
• A stream that breaks midway ends with the API’s error event (Anthropic event: error, OpenAI error chunk). A second generation is never spliced on, and agent tasks are never rerun.

After three restarts

The state becomes failed: <reason>. Requests get a clear 503 that names the reason and points to hearthwork logs --server. hearthwork start clears the failure and tries again.

hearthwork status (and status --json) shows the last failure, the exit code and the last 20 lines of the server log.

See also Common errors and Logs.