The meshmonitor container is running on ubuntu-dev, not docker-dev.
Having both instances fight over the same Meshtastic node TCP connection
was causing repeated ECONNRESET disconnects.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add com.ghost.tel/stack-type labels to all stacks:
- prod (17): Production services from core
- dev-only (11): Experimental/device-specific services
- public (8): Public-facing services (uplink.tel, sequela.tel)
New public stacks from docker-public:
- nitter-public: Nitter instance for uplink.tel
- freshrss-public: FreshRSS for uplink.tel
- rsshub-public: RSSHub for uplink.tel
- searx-public: SearXNG for uplink.tel
- wikijs-public: Wiki.js for sequela.tel
- matomo-public: Matomo analytics for sequela.tel
Also fixes:
- Remove obsolete 'version' key from compose files
- Fix snowflake to remove duplicate watchtower service
- Standardize compose file formatting
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- zerotier: Change exposed port from 3180 to 3000 (actual UI port)
- zerotier: Add loadbalancer.server.port label for traefik
- traefik: Add basicAuth volume mount for dashboard-auth middleware
- test-services.sh: Fix registry URL to use /v2/ endpoint
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
