The github-runner-tts-reader pod was crash-looping (329 restarts) and consuming
memory on the over-pressured old rke2 cluster (rke2-agent1 ~81%), contributing
to Blazor SignalR circuit drops on ttsreader/chat. It provides no working CI in
this state. Set replicas: 0 so ArgoCD stops re-creating it; restore to 1 once
the runner is fixed or CI moves to a working host.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
6950010ea4