I run local backups every night on a proxmox node, and have notifications configured to be sent using a webhook to Pushover. Manual tests of the notification transport always succeed, and the proxmox node also forwards emails from a nightly rsync cron task (that copies my backups offsite). I also have a second proxmox node at the same location, with identical backup and notification configurations, that successfully notifies me each night for both tasks. The nodes do not run their backups concurrently.
If I run the proxmox backup manually I get a notification, but each night when it runs on schedule at 03:00 I get no notification, and an accompanying entry in syslog, eg.
Backups and notifications were working perfectly up until late June 2026, when the notifications started failing sporadically. They fail almost 100% of the time from July 18th onwards. This is similar to an issue posted by @warner_k
This feels like some sort of timing issue. Is it trying to process the notification too soon after the backup finishes? What could the backup job be doing to break the networking in this way?
If I run the proxmox backup manually I get a notification, but each night when it runs on schedule at 03:00 I get no notification, and an accompanying entry in syslog, eg.
Code:
Sep 16 03:00:01 cain pvescheduler[194987]: INFO: starting new backup job: vzdump --mode snapshot --all 1 --storage SSD --quiet 1 --fleecing 0 --notes-template '{{guestname}}' --node cain --compress zstd --prune-backups 'keep-daily=3,keep-monthly=1,keep-weekly=1' --notification-mode notification-system
Sep 16 03:00:01 cain pvescheduler[194987]: INFO: Starting Backup of VM 100 (qemu)
Sep 16 03:01:49 cain pvescheduler[194987]: INFO: Finished Backup of VM 100 (00:01:48)
Sep 16 03:01:49 cain pvescheduler[194987]: INFO: Starting Backup of VM 101 (qemu)
Sep 16 03:05:11 cain pvescheduler[194987]: INFO: Finished Backup of VM 101 (00:03:22)
Sep 16 03:05:11 cain pvescheduler[194987]: INFO: Backup job finished successfully
Sep 16 03:05:12 cain perl[194987]: could not notify via target `Pushover`: io: No route to host (os error 113)Backups and notifications were working perfectly up until late June 2026, when the notifications started failing sporadically. They fail almost 100% of the time from July 18th onwards. This is similar to an issue posted by @warner_k
This feels like some sort of timing issue. Is it trying to process the notification too soon after the backup finishes? What could the backup job be doing to break the networking in this way?