Hey there,
I've been dabbling with Proxmox every once in a while, and one aspect I keep having issues with is the notification system - which is why I'm wondering if I'm missing something, or if this is something that's still being worked on.
The problem I have is that the rule system doesn't appear to be flexible enough, to do basic things I want to do.
For example: most of the notifications come from the backup system. It's great to be informed, but there's so many, so I just want to be informed on warnings and errors for now. So, I go to the backup notification settings in the cluster config, don't select legacy, because who wants to have legacies on a new system, which means I have to use the notification system.
The default notification configuration consists of one rule - send everything to root. So I change the rules to:
"If all rules match: Job-ID = backup-job-id AND severity = warning, error, unknown"
Fair enough. But I still want to be informed of everything else. So I create a new rule, I want to be informed of everything except the backup job for now.
"If all rules match: Job-ID = "
and here I'm stumped already. I can't match "!= Job-ID".
Yeah, I can work with negations, such as "If no rules match: Job-ID = backup-job-id" which will work, but the moment I have any other conditions, it gets really awkward - such as "Job-ID != backup-job-id, and severity == error" - then I have to create rules such as "OK, this should match when the Job-ID isn't backup-id, and also I don't want it to match if the severity is info, warning, debug or unknown". I find that rather contrived (writing effectively "! ( Job-Id == backup--job-id || severity == info,warning,debug,unknown ) " instead of just "Job-Id != backup-id && severity == error") - I order to still understand in 3 years that I wanted to "match everything else if it's an error", I have to add that into the comment field.
Can anyone share their grown rule set of effective notifications?
I've been dabbling with Proxmox every once in a while, and one aspect I keep having issues with is the notification system - which is why I'm wondering if I'm missing something, or if this is something that's still being worked on.
The problem I have is that the rule system doesn't appear to be flexible enough, to do basic things I want to do.
For example: most of the notifications come from the backup system. It's great to be informed, but there's so many, so I just want to be informed on warnings and errors for now. So, I go to the backup notification settings in the cluster config, don't select legacy, because who wants to have legacies on a new system, which means I have to use the notification system.
The default notification configuration consists of one rule - send everything to root. So I change the rules to:
"If all rules match: Job-ID = backup-job-id AND severity = warning, error, unknown"
Fair enough. But I still want to be informed of everything else. So I create a new rule, I want to be informed of everything except the backup job for now.
"If all rules match: Job-ID = "
and here I'm stumped already. I can't match "!= Job-ID".
Yeah, I can work with negations, such as "If no rules match: Job-ID = backup-job-id" which will work, but the moment I have any other conditions, it gets really awkward - such as "Job-ID != backup-job-id, and severity == error" - then I have to create rules such as "OK, this should match when the Job-ID isn't backup-id, and also I don't want it to match if the severity is info, warning, debug or unknown". I find that rather contrived (writing effectively "! ( Job-Id == backup--job-id || severity == info,warning,debug,unknown ) " instead of just "Job-Id != backup-id && severity == error") - I order to still understand in 3 years that I wanted to "match everything else if it's an error", I have to add that into the comment field.
Can anyone share their grown rule set of effective notifications?