По решению владельца: политики на роутере нет, меняем адрес.
БЫЛО: из CT 111 не было выхода в интернет — шлюз не отвечал, при этом
все остальные хосты локальной сети были доступны. Перед этим исключил
маршруты, ARP, политики маршрутизации, правила по IP/MAC/интерфейсу на
хосте, firewall Proxmox (отключён), цепочки Tailscale и второй адрес на
том же интерфейсе.
СТАЛО: 192.168.1.111 -> 192.168.1.240 (в стороне от DHCP-диапазона:
.112, .124, .128 оказались заняты устройствами). Шлюз доступен,
example.com отвечает 200, службы поднялись.
ВМЕСТЕ С АДРЕСОМ поправлены ссылки: настройки контейнера, внешний nginx
на VPS (3), домашний nginx (3), наблюдение (3), реестр обновлений (3),
карта доменов (3).
ОТКРЫТИЕ: домашние домены проксирует nginx НА ХОСТЕ PVE
(/etc/nginx/conf.d/buryans-all.conf, 41 домен), а не NPM в CT 101.
Значит при любой смене адреса искать ссылки надо в двух местах — на VPS
и на хосте. Записано в документ.
Проверено: книги/новости/rss работают дома и снаружи (302),
наблюдение 60 проверок без провалов.
- Xray 26.3.27 with VLESS + TLS + TCP on port 10443
- Let's Encrypt certificate for git.buryans.ru
- NAT configured for outbound traffic routing
- 5 user profiles with unique UUIDs
- Port 10443 opened in provider firewall
- Runbook and audit report added
- Documented final state of Phase 5 (Audit & Transparency)
- Defined infrastructure baseline v1.3 metrics and achievements
- All automation, monitoring, and backup goals met
- Implemented ops/weekly-report.sh to gather 7-day stats (alerts, commits, uptime, LXC)
- Automated Gitea Issue creation via API with 'weekly-report' and 'audit' labels
- Added webhook placeholder for Matrix/Telegram notifications
- Scheduled weekly execution via cron (Sundays at 20:00)
- Implemented ops/audit.sh for sub-second system state snapshotting
- Automated CHANGELOG.md updates with daily metrics (LXC, alerts, disk, git)
- Verified auto-commit and auto-push to Gitea main branch
- Scheduled daily execution via cron at 09:00
- Added supporting scripts to version control (audit.sh, backup.sh)
- Automated mass installation script executed successfully across 28+ LXC containers
- node_exporter (port 9100) deployed and enabled via systemd on all CTs
- postgres_exporter (port 9187) deployed on CT 151 with secure socket auth
- Verified metric exposure: Prometheus now has full observability over all infrastructure
- Fixed JSON parsing to correctly handle n8n webhook payload wrapping (payload.body)
- Verified end-to-end delivery to Telegram with proper HTML formatting
- Grafana Contact Point configured declaratively via provisioning
- Created n8n webhook listener for Prometheus/Grafana alerts
- Implemented smart parsing for both Grafana and Alertmanager payload formats
- Configured Grafana Contact Point declaratively via provisioning
- Alerts are now formatted and sent to Telegram consistently
- Verified network connectivity between Grafana (CT 113) and Prometheus (CT 134)
- Applied declarative datasource provisioning via pct push
- Restarted Grafana to apply configuration
- Datasource is now set as default and ready for dashboards/alerts
- Documented fixes for Unknown Monitor, Telegram duplicates, and queue clearing
- Added SPRINT-F3-STATUS.md for overall tracking
- F3-1, F3-2, F3-3 are now DONE. F3-4 is next.
- Created comprehensive audit script (ops/audit-infrastructure.sh)
- Scanned all LXC containers and nested Docker services
- Generated infrastructure-inventory.json with full service map
- Auto-created runbook templates for all discovered services
Each runbook includes:
- LXC CT ID and Docker container name
- Port mappings and network info
- 5-step procedures (status, health, logs, restart, troubleshooting)
This provides complete operational documentation for the entire Buryan Hub infrastructure.
- Defined architecture (mautrix-telegram)
- Outlined prerequisites, implementation steps, and acceptance criteria
- Documented risks (E2EE, rate limiting) and next steps for future sprint