Sonnet limit hit, can't switch to higher Think model

[jpdupree]

Image Variant

Full (latest / dev)

Image Tag / Version

2.1.87

Host OS

Windows (Docker Desktop / WSL2)

What happened?

Once the Sonnet only usage runs out, Holly Claude tells me I've hit my limit. Pressing the Think button and choosing any option, the Think button goes back to Standard and tells me I've hit my limit again. I can switch the the Shell tab and switch to Opus using /model and it works fine. But the chat does not make it over to the main tab.

What did you expect?

I expected the think button to stay on the option I chose, and use a different model than Sonnet.

Steps to reproduce

1. Get message about hitting limit
2. Click Think button and choose something other than Standard
3. Type Continue and press enter
4. Get message about limit being hit and Think button goes back to Standard.

Docker Compose / Run command

# ==============================================================================
# HolyClaude — Full Configuration
# All options documented inline.
# Detailed docs: https://github.com/CoderLuii/HolyClaude/blob/main/docs/configuration.md
# ==============================================================================

services:
  holyclaude:
    image: coderluii/holyclaude:latest     # Full image (use :slim for smaller download)
    container_name: holyclaude
    hostname: holyclaude
    restart: unless-stopped
    shm_size: 2g                           # Chromium shared memory — increase to 4g for heavy browser use
    network_mode: bridge
    cap_add:
      - SYS_ADMIN                          # Required: Chromium sandboxing
      - SYS_PTRACE                         # Required: debugging tools (strace, lsof)
    security_opt:
      - seccomp=unconfined                 # Required: Chromium syscall requirements
    ports:
      #
      # CloudCLI web UI — this is the only port you need.
      #
      - "3001:3001"
      #
      # Dev server ports — uncomment as needed.
      # These let you access dev servers running inside the container from your host browser.
      #
      # - "3000:3000"                      # Next.js / Express
      # - "4321:4321"                      # Astro
      # - "5173:5173"                      # Vite
      # - "8787:8787"                      # Wrangler (Cloudflare Workers)
      # - "9229:9229"                      # Node.js debugger
    volumes:
      #
      # PERSISTENT DATA
      #
      # ./data/claude — Settings, credentials, API keys, Claude's memory file.
      #                  Survives container rebuilds. NEVER delete this folder.
      #
      - ./data/claude:/home/claude/.claude
      - "D:/Game Dev:/workspace"
      #
      # ./workspace — Your code and projects. Everything you build goes here.
      #               Accessible from your host machine.
      #
      - ./workspace:/workspace
    environment:
      #
      # TIMEZONE
      # Full list: https://en.wikipedia.org/wiki/List_of_tz_database_time_zones
      #
      - TZ=UTC
      #
      # PERFORMANCE
      # Node.js heap memory limit in MB. Increase if you work on large monorepos
      # and hit out-of-memory errors. 4096 (4GB) is a solid default.
      #
      - NODE_OPTIONS=--max-old-space-size=4096
      #
      # USER MAPPING
      # Match these to your host user so files created inside the container
      # have the right ownership on your host. Run `id -u` and `id -g` on your host.
      #
      - PUID=1000
      - PGID=1000
      #
      # SMB/CIFS NETWORK MOUNTS
      # Only enable these if your volumes are on a NAS, Samba share, or CIFS mount.
      # They enable polling-based file watching since network mounts don't support inotify.
      # Leave commented out for local storage — polling uses more CPU.
      #
      # - CHOKIDAR_USEPOLLING=1
      # - WATCHFILES_FORCE_POLLING=true
      #
      # NOTIFICATIONS (optional)
      # Get notified when Claude finishes a task or hits an error.
      # Uses Apprise — supports 100+ services. Also requires creating a flag file
      # inside the container: touch ~/.claude/notify-on
      #
      # - NOTIFY_DISCORD=discord://webhook_id/webhook_token
      # - NOTIFY_TELEGRAM=tg://bot_token/chat_id
      # - NOTIFY_PUSHOVER=pover://user_key@app_token
      # - NOTIFY_SLACK=slack://token_a/token_b/token_c
      # - NOTIFY_EMAIL=mailto://user:pass@gmail.com?to=you@gmail.com
      # - NOTIFY_GOTIFY=gotify://hostname/token
      # - NOTIFY_URLS=                                   # catch-all: comma-separated Apprise URLs
      #
      # AI PROVIDER API KEYS (optional)
      # These are for the OTHER AI CLIs — not Claude Code itself.
      # Claude Code authenticates through the web UI (OAuth or API key).
      # Set these if you want to use Gemini, Codex, or Cursor CLIs.
      #
      # - GEMINI_API_KEY=your_key
      # - OPENAI_API_KEY=your_key
      # - CURSOR_API_KEY=your_key

Logs / Error output

[CoderLuii]

Thanks. I do think there was a real state problem here.

What I fixed in v1.1.9 is the part I could verify cleanly from the current code path: the selected thinking mode in the main chat flow now persists instead of resetting after each send.

I am leaving this issue open because that is not enough evidence yet to claim the full shell and main-chat model-sync problem is completely solved under the exact rate-limit scenario you described.

So this is partially addressed, but not closed on faith.

If v1.1.9 fixes the behavior for you, I’ll close it. If it still snaps back or loses the selected model, reply here and I’ll keep going.

[jpdupree]

Awesome, thank you! Once I'm able to replicate the scenario I'll let you know.