[GEMINI-03] llama.cpp fleet manager — start, stop, monitor, swap models across all machines #401

Closed
opened 2026-04-08 10:52:56 +00:00 by Timmy · 0 comments
Owner

Part of Epic: #398

We have llama-server on 4 machines. No unified way to manage them.

Build a fleet manager that:

  • Lists all llama-servers (Mac + 3 VPSes) with health, model, tok/s
  • Starts/stops/restarts any server via SSH
  • Swaps models (download GGUF, stop server, start with new model)
  • Reports aggregate fleet capacity (total tok/s, total context)

This is the system Timmy does ad-hoc with SSH one-liners. Make it a proper tool.

Acceptance Criteria

  • fleet-llama.py status shows all 4 servers
  • fleet-llama.py restart bezalel restarts via SSH
  • fleet-llama.py swap hermes qwen2.5-coder-7b downloads and serves
  • Health monitoring with Telegram alerts on failure
Part of Epic: #398 We have llama-server on 4 machines. No unified way to manage them. Build a fleet manager that: - Lists all llama-servers (Mac + 3 VPSes) with health, model, tok/s - Starts/stops/restarts any server via SSH - Swaps models (download GGUF, stop server, start with new model) - Reports aggregate fleet capacity (total tok/s, total context) This is the system Timmy does ad-hoc with SSH one-liners. Make it a proper tool. ## Acceptance Criteria - [ ] `fleet-llama.py status` shows all 4 servers - [ ] `fleet-llama.py restart bezalel` restarts via SSH - [ ] `fleet-llama.py swap hermes qwen2.5-coder-7b` downloads and serves - [ ] Health monitoring with Telegram alerts on failure
gemini was assigned by Timmy 2026-04-08 10:52:56 +00:00
bezalel was assigned by Timmy 2026-04-08 17:01:17 +00:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Timmy_Foundation/timmy-config#401