Prometheus¶
LXC: monitoring (CT 202) - metrics collector.
Instance¶
| Setting | Value |
|---|---|
| Port | 9090 |
| Config | /data/prometheus.yml (bind-mounted) |
| URL | https://prometheus.home.example.com:<admin-port> |
| Managed as | Komodo stack monitoring |
Scrape Targets¶
Job: prometheus
- localhost:9090 - self
Job: pve
- pve-exporter:9221 (metrics_path /pve, target param 192.168.50.112) - Proxmox host
Job: node - single job, instance label per target:
| Instance | Target |
|---|---|
management |
192.168.50.100:9100 |
pihole |
192.168.50.101:9100 |
ultrafeeder |
192.168.50.200:9100 |
iot |
192.168.50.201:9100 |
media-arr |
192.168.50.30:9100 |
media-server |
192.168.50.31:9100 |
media-dl |
192.168.50.32:9100 |
game-panel |
192.168.50.40:9100 |
minecraft-wings |
192.168.50.41:9100 |
monitoring |
192.168.50.202:9100 |
CT 600 is not present - no node-exporter installed.
Job: ultrafeeder
- 192.168.50.200:9273 - ADS-B stats (requires :telegraf image tag)
- 192.168.50.200:9274 - feed stats
Config Reload¶
Editing prometheus.yml¶
Edit in-place only - editors that do atomic rename-based saves (write to temp file, then rename) replace the file's inode; the running Prometheus container stays bound to the old one and silently keeps serving the previous config.