content improve 2

infra/index.njk

@@ -36,3 +36,15 @@ sections:
 <p class="site-intro">A peek under the hood of my homelab. Proxmox, Synology, Docker, and how everything is wired together.</p>
 
 {% include "card-grid.njk" %}
+
+<section class="about-expanded">
+  <h2>The setup at a glance</h2>
+  <p>The infrastructure runs on three main machines: a Proxmox VE host (Intel NUC, 64GB RAM) as the primary compute node, a Synology DS918+ NAS for storage and Docker services, and a Xubuntu VM on Proxmox for development. An OpenWrt router handles multi-WAN failover with WiFi and Powerline backhaul, and four Zyxel GS1200-8HP managed switches distribute the network across the flat.</p>
+  <p>The whole stack is monitored via Prometheus and Grafana, with node exporters on every machine, cAdvisor for container metrics, and SNMP exporters for network hardware. Backups follow a 3-2-1 strategy: local snapshots, NAS copies, offsite sync to a Raspberry Pi in a different room, and cloud archive for worst-case scenarios.</p>
+
+  <h2>Why Proxmox</h2>
+  <p>Proxmox VE was chosen over alternatives like ESXi or plain Docker for several reasons. The open-source model means no licensing headaches. LXC containers provide near-native performance for lightweight services without the overhead of full VMs. The web UI makes it easy to manage snapshots, backups, and migrations. And the ZFS support on the underlying Debian system gives us reliable snapshots and send/receive for the backup chain.</p>
+
+  <h2>How services are organized</h2>
+  <p>The general rule: stateful services (databases, media archives, monitoring data) run on the Synology NAS where they benefit from the integrated backup and snapshot system. Stateless services and compute-heavy workloads run on the Proxmox Docker VM. LXC containers handle lightweight always-on services like Pi-hole and MariaDB. This separation makes it easy to tear down and recreate Docker containers without worrying about data loss.</p>
+</section>

infra/monitoring/index.njk

@@ -9,7 +9,9 @@ layout: base.njk
 <div class="detail-content">
 
 <p>Prometheus and Grafana form the central monitoring stack for the whole
-homelab, running as Docker containers on the NAS.</p>
+homelab, running as Docker containers on the NAS. Every machine in the
+flat — Proxmox host, Docker VM, Synology NAS, and Raspberry Pi devices —
+feeds metrics into this central instance.</p>
 
 <h2>What's monitored</h2>
 <ul>
@@ -27,6 +29,23 @@ homelab, running as Docker containers on the NAS.</p>
 <tr><td>Network printer</td><td>Toner level, page count, online/offline status via SNMP</td></tr>
 </table>
 
+<h2>How it's deployed</h2>
+<p>Prometheus runs as a Docker container on the Synology NAS with a
+bind-mounted data directory for persistence. Grafana is deployed as a
+sibling container, connected to Prometheus as its primary datasource.
+Both containers are defined in a single Docker Compose file and managed
+via Portainer. The Prometheus configuration file defines scrape targets,
+scrape intervals (15 seconds for hosts, 60 seconds for network devices),
+and retention policies.</p>
+
+<h2>Alerting</h2>
+<p>Basic alerting is configured through Prometheus Alertmanager, which
+sends notifications to a dedicated Telegram channel for critical events
+like host-down or disk-full conditions. The dead-man's-switch pattern
+ensures that if Prometheus itself stops scraping, an alert fires within
+minutes. More sophisticated alerting rules (disk space prediction,
+temperature thresholds) are planned but not yet implemented.</p>
+
 <h2>Open items</h2>
 <ul>
 <li>Alerting rules are not yet configured for most services.</li>