Best practices of monitoring in medium/large enviroments

[Sypher]

Hi,

We're running Zabbix on a few hunderd hosts and started thinking about how to improve it.

At the moment we are using the "Overview" page with "Group: All" and "Type: Triggers" at "Zabbix fullscreen" However, we're using quite some triggers and the amount of hosts is growing weekly. Loading this page takes quite some time, but the other problem is that we have to scroll like maniacs in order to keep a close eye out to the hosts' and trigger's status.

I'm looking for "best practices" from my fellow Zabbix users who are monitoring medium/large enviroments (starting at 100 hosts I guess?).

What are you using to monitor your hosts?
Are you using the Dashboard, Overview, The latest data overview, the list of triggered triggers or just mail notifications? Are you using a custom frontend or custom MySQL queries?

I'm looking forward to your responses!

[tchjts1]

We have been using Zabbix for 2 years now. I think I have looked at the "Overview" page twice in that time. IMO, that particular page is one that could go away.

I find all the information I need from the dashboard on what issues are currently in progress. No scrolling needed unless you have a ton of host groups defined. You can simply hover your mouse over the number that is in the severity column to see exactly what the problem is on what server and how long it has been an issue.

[gospodin.horoshiy]

I rely on Dashboard, checking triggers status and checking emails those triggers might generate. Also, we've been using some screens that are the same for each OS type and contains all hardware graphs so we can analyze its usage. Overview, Triggers and Events sections are not that informational.

[richlv]

personally, i find monitoring -> triggers page to be one of the best locations.
you can set minimal severity to b shown in the filter, and that way you can see all the currently active problems of some higher severities.

dashboard with the new filtering options also should now be quite useful in providing high level overview about most important hostgroups

[tchjts1]

personally, i find monitoring -> triggers page to be one of the best locations.

I hate those beeps every time you go to that page. It is like fingernails across a chalkboard to me.

[richlv]

I hate those beeps every time you go to that page. It is like fingernails across a chalkboard to me.

hmm. did you try clicking the button in the upper right corner, the one that looks like a loudspeaker ?

[tchjts1]

hmm. did you try clicking the button in the upper right corner, the one that looks like a loudspeaker ?

Oh geez no! Why would I want to make it louder??

[tchjts1]

hmm. did you try clicking the button in the upper right corner, the one that looks like a loudspeaker ?

In seriousness... that feature does not work as expected. I went to the triggers page, disabled the sounds, acknowledged some alerts through "acknowledge and return" ... when I was returned to the page the icon showed sounds were turned on and the beeping began again.

If this were a page/function I had any intention of using, I would file a formal bug alert... but since I don't, this is the extent of my bug report.

[nelsonab]

I hate those beeps every time you go to that page. It is like fingernails across a chalkboard to me.

First time I heard it back in the 1.4 series it was late at night and I was wondering what the heck my alarm clock was doing going off, and for that matter what the hell was it doing in the kitchen?

[Aly]

First time I heard it back in the 1.4 series it was late at night and I was wondering what the heck my alarm clock was doing going off, and for that matter what the hell was it doing in the kitchen?

Robbing refrigerator?!

P.S. Fixed mute alarm button.

[gospodin.horoshiy]

First time I heard it back in the 1.4 series it was late at night and I was wondering what the heck my alarm clock was doing going off, and for that matter what the hell was it doing in the kitchen?

Haha, making it sound like default alarm clock was a cruel joke, devs!

[Kerrygeek]

Yeah, it took a little getting used to but it only beeps when I first start it in the morning. I leave the triggers page up all the time, usually in the background, so when it beeps I know to check it. Our help desk uses the overview.

Kerry

[Sypher]

Previously I was using the Overview page but since there apparently is a fix for showing only items with "Problem" status, I decided to give this a go.

A list with all problems is a lot more practical to use than the overview.
Since we've got quite an amount of servers to monitor, items got overlooked because of that.

[radamand]

We have around ~300 hosts being monitored and mostly rely on email notifications.
If a notification comes in and I need to look at latest data or such I can, but only as needed.

[Alexei]

Haha, making it sound like default alarm clock was a cruel joke, devs!

Thanks!