Zabbix template

[vashvampir]

Daily reset of the failed tasks to 0 at 00:00 should be fine as long as the alerts won't get autocleared when it drops to 0 again. So that users will even notice the failed jobs when not logged in to zabbix webUI for 24 hours.

Ok, give me some time to think and experiment with the PBS API, I'll try to figure out how it would be the most logical way to monitor

 

[itNGO]

Created new template - Proxmox Backup Server by HTTP. https://github.com/ismvru/zbx-tmplt-pbs

Sample latest data:

View attachment 40480

Nice! But can it monitor Backups? Or is there a way to get a trigger if a Backup has failed?

 

[Dunuin]

Nice! But can it monitor Backups? Or is there a way to get a trigger if a Backup has failed?

Sort of. He is working on it. See the last 6 posts.

 

[vashvampir]

Sort of. He is working on it. See the last 6 posts.

Nice! But can it monitor Backups? Or is there a way to get a trigger if a Backup has failed?

Updated template:

1. Item "PBS: New failed tasks" has been removed. In the future, if necessary, new items will be added according to the number of failed tasks of a certain type.
2. Added new items:

- PBS: Failed tasks - Failed tasks for all time (Without history)
- PBS: Failed backup tasks count - Failed backup tasks count for all time
- PBS: Last failed backup task - Last failed backup task data (In order to prevent the zabbix database from growing, it discards new values if the value received from PBS has not changed)
- PBS: Last failed backup task time - Last failed backup task finish time (in unixtime format)

3. New trigger added - PBS: New failed backup task
Fires every time when a new backup job has failed (item "PBS: Last failed backup task" was changed) in the last 24 hours
Allows only manual close

UPD
Added recovery expression to PBS: New failed backup task. Closes if no backup job has failed in the last 24 hours.

 

[vashvampir]

Oops, I think I found a bug. The trigger fires only once, and fires again if closed by hands. Thinking how to fix...

Upd.
The trigger can now be closed 5 minutes after it fires. Now the values of the item "PBS: Last failed backup task" are saved once every 5 minutes in the history if the value on the PBS server has not changed, and once per minute if the value changes.

 

[Dunuin]

I think other tasks would also be important. Not just failed backup tasks. For example a sync task where you would be without a second offsite backup. Or a verify task that found corrupted or missing chunks.
But not sure if a verify task will return a error if it found bad chunks of it would be required to verify the state of every single backup snapshot to find out if there are any corrupted backups.

 

[vashvampir]

I think other tasks would also be important. Not just failed backup tasks. For example a sync task where you would be without a second offsite backup. Or a verify task that found corrupted or missing chunks.
But not sure if a verify task will return a error if it found bad chunks of it would be required to verify the state of every single backup snapshot to find out if there are any corrupted backups.

Ok, I will return the trigger when there are new failed tasks.
It remains to understand how to do it correctly, because I will have to start not from the number of failed tasks, but from changing the response from the API (the last failed task).