Notification issues

[sahartford]

I've configured the email notifications for a number of things including backups and host backups and the daily digest. Those I'm not getting although test works and I do get alerts on things like utilization, invalid logins, etc.

[MacRimi]

Hi, thank you for reporting this.

Those three do not travel the same way as the alerts you are receiving, which is why the test and the usage alerts work while they don't:

• Host backups: I found a bug here. With HTTPS enabled in the Monitor, the host backup could not reach it to send its notification. This is fixed in the develop branch. Do you have HTTPS enabled in the Monitor?
• VM and CT backups: these come from Proxmox itself, which calls the Monitor when a job ends. If the backup job has an address in "Send email to", Proxmox may send that mail on its own and skip the Monitor. Could you check whether your job has one, and what Datacenter → Backup → your job → Notification mode says?
• Daily digest: it only groups informational events, and backups are not among them because they are sent right away. When there is nothing to group, no email is sent.

It would also help to know your ProxMenux version, and whether Settings → Notifications → History shows any backup entry.

[sahartford]

1 - For the host backups, I am not using HTTPS in Monitor. Also confirmed they run and complete each night
2 - We resolved this, was incorrectly configured. Now get both the notification from Proxmox and ProxMenuX
3 - Yes, that was my understanding, info events, but I would assume over the course of a day there would be any number of info events? I guess maybe not in relation to what ProxMenuX might be tracking through...

Running version 1.2.6.
In System Logs and Events -> Notifications - does show the "success" and "task-log" for the backup task.

[MacRimi]

Hi @sahartford, thanks for checking those, and sorry for the wait.

Host backups. You were right that HTTPS wasn't your case. Going through it again I found the actual reason. The Monitor treated "host backup complete" as the same event every time and left 24 hours between two identical notifications, so a job that runs every night kept falling inside that window. On top of that, on a channel with the daily digest enabled the result was kept for the digest instead of being sent when the job ended. What I told you before, that backups are sent right away, was true for VM and CT backups but not for host backups.

Both are corrected in the development branch: every run of a host backup is now sent when it ends, like a VM or CT backup. It will be in the 1.2.7 version.

VM and CT backups. Good to hear that one is sorted.

Daily digest. It is only sent on days when at least one informational event was held for it, for example an LXC update that completed. Warnings, errors, backups and VM or CT starts and stops are sent live and never go into it, so on a quiet host there can be several days with nothing to summarise, and then no email is sent. Each run is recorded, though: Settings → Notifications → History shows a digest entry for every day, saying whether it had something to send. If those entries are there, the schedule is working.

[sahartford]

Thanks for the update.

Re: Daily Digest, you mention checking "Settings -> Notifications -> History". I don't have/see that section.

[MacRimi]

Hi @sahartford, sorry, I told you it would be there in the next version, but it didn't make it into this beta. It will be available in the next stable version, which is coming out very soon.

If you want to check it now, you can do it from the host's shell. This only reads the history, it doesn't change anything:

python3 -c "import sqlite3; c = sqlite3.connect('file:/usr/local/share/proxmenux/health_monitor.db?mode=ro', uri=True); [print(*r) for r in c.execute("SELECT sent_at, channel, success, message FROM notification_history WHERE event_type = 'digest' ORDER BY id DESC LIMIT 10")]"

Each line is one day's digest run for a channel. If you see a line for each day, the schedule is working as it should.