SyslogSolution/prose-contracts
7
0
Fork 0
Code Issues 2 Pull Requests Actions Packages Projects Releases Wiki Activity

docs: add PBS GC schedule to proxmox-monitor contract
PR Pipeline — Authorize → Validate → Review → Merge / auth (pull_request) Successful in 5s
Details
PR Pipeline — Authorize → Validate → Review → Merge / validate (pull_request) Successful in 4s
Details
PR Pipeline — Authorize → Validate → Review → Merge / lint (pull_request) Successful in 2s
Details
PR Pipeline — Authorize → Validate → Review → Merge / ai-review (pull_request) Successful in 12s
Details
PR Pipeline — Authorize → Validate → Review → Merge / gate (pull_request) Successful in 0s
Details

Browse Source 
- GC runs daily at 20:00 UTC (after backup window closes at 06:40Z)
- Volume is /tank/pbs-backup on ZFS pool 'tank' (12.7T, 10% used)
- NOT /media/easystore2 (media library, 3.7T, 96% HOST-RED)
- Cron: /etc/cron.d/pbs-gc on storepve (192.168.68.6)
- Last run: 2026-09-18 00:08:34 UTC (removed 551.768 GiB)
This commit is contained in:
root
2026-09-18 04:46:44 +00:00
parent 1137dd4582
commit 1bfab85a25
1 changed files with 18 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
proxmox-monitor.prose.md
+18
View File
@@ -142,3 +142,21 @@ each probe. If any probe returns non-200, flag as alert.
- **PVE exporter metric schema**: NOT name-prefixed. `pve_cpu_usage_ratio`, `pve_memory_usage_bytes`, `pve_disk_usage_bytes`, `pve_uptime_seconds` are GUEST-level only (24 series, `id=lxc/100` etc). Node-level host metrics come from node_exporter. Storage pool usage: `pve_storage_info` (info only, no usage bytes — use node_filesystem_* for actual disk usage).
- **grafana piechart plugin removed** from `GF_INSTALL_PLUGINS` (Angular, unsupported in Grafana 13).
- **Single pve-exporter points at amdpve .15** — if amdpve API is down, cluster metrics gap (other node_exporters still report host metrics). Acceptable; amdpve is primary.
## PBS Garbage Collection Schedule (storepve-datastore)
**Schedule:** Daily at 20:00 UTC (4:00 PM EDT)
**Rationale:** The nightly backup window runs 04:00–06:40 UTC (local backups 04:00–06:40Z, S3 sync 04:15Z, S3 trim 05:15Z). Running GC during this window causes avoidable I/O contention on the same datastore and host. 20:00 UTC lands well after the backup window closes and before the next day's backups begin.
**Volume:** `/tank/pbs-backup` on the ZFS pool `tank` (12.7T total, 11.3T free, 10% used). Inside CT 107, this appears as 1.5T total / 413G used / 1.1T avail = 28%.
**Important:** This is NOT the same as `/media/easystore2` (3.7T, 96% used, HOST-RED), which is a media library ("4K MOVIES") on the storepve host root filesystem. The PBS datastore lives on a separate ZFS pool.
**Cron:** `/etc/cron.d/pbs-gc` on storepve (192.168.68.6):
```
0 20 * * * root /usr/local/bin/pbs-gc.sh
```
**Last GC run:** 2026-09-18 00:08:34 UTC (3h 19m 9s, removed 551.768 GiB)
**Schedule owner:** The `gc-schedule` field in the PBS datastore config is the durable owner of GC behavior. The cron is a fallback because `proxmox-backup-manager datastore update --gc-schedule` failed to parse the calendar event (Nom(Eof) error).